Status
Current implementation report for IARPG-OPS-2 2.0.17-wip. The Campaign Branch Archive is a fictional, browser-local, deterministic comparison prototype. Its evidence establishes repository and harness behavior only; it does not establish geopolitical truth, entertainment quality, human comprehension, or production readiness.
Purpose
Document the extension of the Campaign State Console into a multi-record archive that can import, validate, replay, compare, quarantine, and export fictional campaign branches while preserving source bytes, unknown extensions, neutral branch identity, correction history, and reversible return points.
Scope
This report covers /campaign-archive, assets/js/campaign-archive.js, the shared deterministic model in scripts/campaign_model.py, two published campaign-record examples, the comparison example and schema, six additional safe campaign briefs, campaign and record schema changes, validation evidence, discovery, and UAI routing. It does not authorize real-world monitoring, profiling, recruitment, targeting, covert contact, intrusion, evasion, sabotage, or weapons activity.
Executive Summary
IARPG now supports deterministic comparison of two or more fictional campaign histories that share the immutable 2.0.16-wip relationship baseline. An imported record is evaluated against known contract fields and replayed turn by turn. Unknown extension values remain in the original record; invalid records are quarantined and can be downloaded in the exact imported byte sequence rather than silently repaired or discarded.
The archive assigns neutral aliases such as Branch Alpha and Branch Bravo by import order. It does not derive branch names from filenames, player names, organizations, countries, protected characteristics, or inferred identity. Designers can select any committed turn—including Turn 0—as a replay return point. The interface then compares versioned relationship states, domain postures, burdens, dependencies, consultation status, evidence confidence, community consequences, reason codes, turn decisions, evidence events, correction paths, and reversal availability.
The compact comparison export contains only the branch identifiers and deterministic comparison record needed for portable review. It does not embed the complete source bodies or invent a universal friend/enemy, loyalty, morality, trustworthiness, threat, dangerousness, or recruitability score.
Evidence Reviewed
- [Deterministic Alliance Campaign State and Mission Expansion Report](deterministic-alliance-campaign-state-and-mission-expansion-report.md#findings)
- [Alliance and Recruitment Worldbuilding System](alliance-and-recruitment-worldbuilding-system.md#reviewed-synthesis)
- [How Alliances Actually Work](how-alliances-actually-work.md#reviewed-synthesis)
- [Faction Archetypes and Recruitment Ecosystems](faction-archetypes-and-recruitment-ecosystems.md#reviewed-synthesis)
- [Lone-Actor Mobilization, Network Influence, and Disengagement](lone-actor-mobilization-network-influence-and-disengagement.md#reviewed-synthesis)
- Campaign-state publication
- Campaign-record schema
- Campaign comparison schema
- Published branch A
- Published branch B
- Published comparison
Findings
The archive replays rather than trusts imported summaries
Each valid import must identify the published campaign-record schema, the IARPG release family, the immutable campaign ID, and baseline version 2.0.16-wip. Turns must be sequential. For every imported turn, the archive reconstructs the expected outcome from the published scenario and response identifiers, compares all known turn fields, then applies the expected relationship versions. At the maximum return point, the imported current-state array must also match the deterministic reconstruction across all 27 relationships.
This prevents a record from claiming a state that its own turn history does not produce. It also keeps the authoritative relationship transition logic in the campaign publication rather than in user-supplied summaries.
Original bytes and extension values are preserved
The browser reads every selected file into a Uint8Array before parsing. That original byte sequence remains available for local download whether validation succeeds or fails. Validation projects only the known contract fields for comparison. Unknown extension paths are listed for reviewers, while their values remain intact inside the preserved original.
The archive does not normalize whitespace, rewrite JSON ordering, remove extensions, repair broken fields, or overwrite the imported file. A malformed or non-replayable branch is visibly quarantined and excluded from comparison.
Neutral aliases prevent accidental identity signaling
Branches receive generated aliases by import position. The displayed alias and compact comparison ID do not contain the source filename, record author, player identity, organization, geography, or protected characteristic. Removing a branch regenerates the remaining neutral sequence. This is a usability and fairness safeguard, not anonymization of external metadata outside the browser.
Return points make corrections and reversals inspectable
Every valid branch exposes Turn 0 through its maximum committed turn as selectable return points. Selecting a return point reconstructs that branch from the immutable baseline plus only the turns through the selected point. The comparison can therefore answer not only “Where did these histories end?” but also “At which decision did their institutional states begin to diverge?”
The timeline exposes scenario, response, reason codes, evidence events, consultation decisions, correction actions, mission references, and the range of reversible return points. No prior state is deleted to make a later branch appear cleaner.
Relationship comparison remains domain-specific
The archive compares each relationship’s lifecycle stage, version, effective turn, domain postures, burden distribution, bilateral dependencies, consultation state, evidence confidence, community impact, reason code, evidence-event references, reversibility, and superseded version. A relationship may match in one domain and differ in another.
The design explicitly rejects one-dimensional alignment. It does not calculate a universal friendship, hostility, loyalty, morality, trustworthiness, threat, dangerousness, recruitability, or targetability score.
Turn comparison preserves institutional reasons and consequences
When branches differ at the same turn position, the archive shows each branch’s stress event, selected response, reason codes, evidence-event IDs, consultation decision, correction path, and mission references. This keeps the explanation attached to the state change rather than presenting unexplained color-coded divergence.
Affected-community consequences and internal-bloc responses remain inside the source campaign turns and deterministic replay. They are not reduced to demographic prediction or identity-based suspicion.
The browser-local boundary is explicit and testable
The archive has no account, remote API, analytics event, fetch, XMLHttpRequest, WebSocket, sendBeacon, localStorage, or sessionStorage workflow. It asks for no personal identity field. Imports and comparison records exist only in the active tab unless the visitor explicitly downloads an original or exports the compact comparison.
The page publishes runtime counters declaring zero network requests, persistent-storage writes, and personal-identity fields. Managed Chromium validation independently instruments those interfaces and confirms zero use during the tested workflow.
Six new briefs make branch comparison narratively useful
The campaign publication now contains 18 safe institutional mission briefs. The six new briefs focus on coalition mediation, relief coordination, oversight disputes, mandate change, lawful institutional refusal, and reintegration. They are designed to produce different but reviewable campaign histories through consultation, evidence, remedy, and mandate choices rather than real-world wrongdoing instructions.
Decisions or Recommendations
- Keep the 2.0.16-wip 27-relationship baseline immutable for every record accepted by this archive generation.
- Treat deterministic replay as the acceptance gate; do not trust imported counts or current-state summaries without reconstruction.
- Preserve the complete original bytes for both valid and invalid imports until the active browser workspace is cleared.
- Preserve unknown extension fields in source records and list only their paths in compact comparison output.
- Keep branch aliases neutral and independent of filenames, authors, players, countries, organizations, diagnoses, or protected characteristics.
- Require future comparison dimensions to remain domain-specific, evidence-linked, reason-coded, correction-aware, and reversible.
- Do not introduce aggregate loyalty, morality, threat, dangerousness, recruitability, or targetability scoring.
- A future round should add a browser-local campaign package assembler that can select reviewed turns from compared branches and create a new proposed branch without mutating any imported source.
Risks and Limitations
- Deterministic equality proves implementation consistency, not that the campaign choices are balanced, complete, or entertaining.
- The browser retains original bytes only while the page is open; it is not a durable archival service.
- Neutral aliases prevent interface identity cues but do not erase metadata that may exist in an imported file outside the known contract.
- Unknown extension fields are preserved but not semantically validated by the current release.
- The comparison schema is a version 1.0.0 publication and may require additive extension as more branch-analysis needs emerge.
- Managed Chromium fixture evidence is not unrestricted external-host navigation, native Firefox, or native assistive-technology speech output.
- Real-world alliance and faction claims in durable research remain subject to their review status and current authoritative verification requirements.
Validation Performed
- Validated the campaign publication, alliance network, both branch examples, and the comparison example against local JSON Schemas.
- Replayed both four-turn example histories from the immutable 2.0.16-wip baseline and reconstructed all 27 relationship states.
- Rebuilt the published comparison independently and matched 2 branches, 12 divergent relationships, 8 divergent domains, 4 divergent turn positions, 12 referenced evidence events, and 8 visible correction paths.
- Confirmed unknown extension paths remain represented and values remain in the original example records.
- Mutated a known response identifier, confirmed deterministic rejection, and confirmed the source object remained byte-equivalent after validation.
- Exercised example loading, neutral aliases, return-point selection, comparison refresh, compact export, invalid-file quarantine, byte-identical invalid-source download, desktop and 320-pixel containment, and the sole sticky
.menu-strip. - Instrumented browser network and persistence interfaces; observed zero runtime network requests, zero persistent-storage writes, and zero personal-identity inputs.
- Re-ran the repository’s static, schema, source-integrity, memory-link, accessibility-structure, browser, host, discovery, checksum, finalization-idempotency, archive-integrity, and fresh-extraction suites before packaging.
Memory References
- Campaign archive hot memory
- Campaign-state hot memory
- Alliance-network hot memory
- Architecture hot memory
- Progress hot memory
- Test-plan hot memory
- Long-term-memory pointer ledger
Related Durable Documents
- [Alliance, Faction, and Network World-System Integration Report](alliance-faction-and-network-world-system-integration-report.md#findings)
- [Interface Density and Faction Hero Report](interface-density-and-navigation-polish-report.md#findings)
- [UAI Memory Organization Report](uai-memory-organization-report.md#findings)
- [International Intelligence Fairness and Coverage Framework](equal-standards-across-unequal-records-international-institutional-fairness-rights-and-evidence-audit.md#bias-audit-terminology-crosswalk-and-editorial-rules)
Supersession Status
Current. This report augments rather than replaces the campaign-state implementation report. It fulfills that report’s recommendation for deterministic branch comparison and replay while leaving the immutable baseline and current campaign-turn contract in force.
