1. Collect
We gather books, articles, archives, reports, and reference works relevant to a chapter, person, or topic.
❦
Public Trust · Editorial Method
❦Methodology
✦
How Bengal Unfolded handles claims, sources, memory, and editorial judgment.
This project is designed as a bilingual public-history reference, not a fast-moving news surface. The methodology below explains how chapters are written, how evidence is classified, and how uncertainty is handled.
Est. 1947 · Bengal✺A Bilingual Archive
What This Page Covers
✦
The minimum rules readers should expect across events, figures, books, resources, and topic hubs.
- Every historical page should connect claims to explicit sources whenever the model supports it.
- Source quality labels describe evidence class, not automatic truthfulness.
- Sensitive or contested history is written with neutrality, caution, and visible sourcing.
- English and Bangla should carry the same factual meaning, not separate narratives.
Editorial Workflow
✦
From source collection to public page publication.
2. Structure
Claims are mapped into normalized content fields so events, figures, periods, places, and resources remain linkable.
3. Cite
Timeline items and major narrative sections use explicit source IDs and evidence-level metadata where required.
4. Review
Sensitive framing, wording, localization parity, and linked-entity integrity are checked before publication.
Corrections and Revisions
✦
This is a living archive. Accuracy work continues after publication.
When a claim, label, or linkage is found to be weak, incomplete, or wrong, the expected response is correction with traceable updates. Major feature and model changes are recorded in the changelog, and content-model changes are expected to remain synchronized with validation and documentation.
Where This Leads
This page is a public-facing foundation for later trust features such as contested-history markers, debate blocks, version history, and correction workflows.
Source Quality Classes
✦
These labels indicate what kind of source a reader is looking at.
Primary
Original or first-hand material such as official records, direct testimony, speeches, or first-party documents.
Secondary
Analytical or reference-oriented works that interpret, synthesize, or explain historical material.
Archive
Collections, repositories, and catalog-style archival surfaces that preserve or expose historical records.
Editorial
Narrative, interpretive, or opinion-driven works where persuasion or storytelling is stronger than documentary method.
Evidence Levels
✦
Evidence labels reflect confidence in support for a specific claim, not a claim’s moral or political importance.
High
Used when a claim is strongly supported by reliable and specific source material.
Medium
Used when support is meaningful but partial, indirect, or dependent on synthesis across sources.
Low
Used when support is limited, interpretive, or still useful mainly as contextual guidance rather than firm confirmation.
Contested and Sensitive History
✦
Not all historical memory is settled. The writing standard is restraint, not rhetorical victory.
- When disagreement exists in sources, the preferred approach is to summarize the disagreement rather than flatten it into one absolute claim.
- Potentially traumatic, violent, or politically sensitive chapters should use stronger sourcing and may carry reader-facing warnings.
- Source quality and evidence level are shown so readers can distinguish well-supported claims from weaker contextual framing.
Bilingual Parity
✦
Localization is not allowed to create two different factual histories.
- English and Bangla should preserve the same core factual meaning.
- If one locale changes factual content, the other locale should be updated in the same change set.
- Proper nouns, dates, and linked entity references should stay aligned across locales.
Related Internal Docs
✦
The public page is backed by stricter internal rules used during editing and validation.
AI ContractEditorial RulesContent ModelSource Quality Rubric