Thesis
The Birthday Book scan can be published as a source object while keeping attribution, identity, handwriting, OCR, and photo interpretation under separate review controls.
The Birthday Book route is not just another document viewer. It is an editorial test of how to publish a source object that invites overreading. OCR is noisy. Handwriting can be uncertain. A table-of-contents string is not the same thing as a verified contribution. The reader's job is to preserve the scan and slow down attribution.
Key findings
The Birthday Book route is a public evidence reader with source provenance and GAPS labels.
The reader separates page images, public transcription rules, contributor status, review queue, and methodology.
The archive states that OCR, handwriting, signatures, and photographs do not by themselves verify identity or conduct.
Unresolved attribution remains a review queue rather than a claim.
Primary-source chain
Claim -> Source -> Bias -> Silence -> Confidence
Claim: The Birthday Book scan can be published as a source object while keeping attribution, identity, handwriting, OCR, and photo interpretation under separate review controls.
Source: the linked archive routes below are the source chain for this page.
Bias: this page is an archive-created editorial reading of public source routes. It favors source visibility and may understate interpretations that lack direct receipts.
Silence: anything not supported by the links below remains an unresolved question, not a hidden fact.
Confidence: moderate to high for route existence and narrow source descriptions; claim-specific confidence remains tied to each linked receipt.
Timeline
- The scan becomes a public source object.
- Page images, OCR gates, review cases, and identity records are structured.
- The route publishes conservative attribution language and GAPS labels.
What could be read another way
An alternate approach would either publish every OCR string as searchable text or suppress the whole object because attribution is risky. The current route is stronger because it shows the source while separating confidence layers.
What this page does not prove
- This page does not verify every contribution.
- It does not identify people from photographs.
- It does not treat OCR as a final transcript.
- It does not defeat redactions or publish unsafe private data.
What still needs source work
- Which pages have public-quality transcription?
- Which attribution records need external source support?
- Which unresolved cases should be prioritized for human review?
Reader value
A generic archive would dump a PDF. An evidence newsroom makes the reading rules visible before the reader reaches for a conclusion.
Evidence links