A Documentary Source Pack Is Not a Bibliography

A bibliography answers one question: what material was consulted?

A documentary source pack has to answer several harder questions. What can this film claim? What remains disputed? Which objects can inform the visuals? Which assets can actually be reused? Who reviewed the evidence before the expensive production stages began?

That is why a list of links is not a production-ready source pack.

The question determines the pack

Research without a bounded question expands forever.

“The Roman Empire” can produce thousands of credible sources and no clear film. “How grain supply shaped daily life in Rome during one period” creates a manageable geography, timeline, vocabulary, and set of factual risks.

The pack should begin with the central question, working thesis, target viewer, time and place boundary, exclusions, and claims that would materially change the film if wrong.

This front page turns research into an editorial contract. It also helps a reviewer reject attractive material that belongs to another episode.

Every source needs a job

For each source, record creator or institution, title, publication or record date, stable locator, access date, source type, relevant pages or timestamps, and a plain statement of what it supports.

Also record what it does not support.

A diary can document what one person perceived. It may not establish how everyone experienced an event. An institutional report can establish what the institution recorded. It may not prove that the report was complete or unbiased. Later scholarship can provide context while disagreeing about causality.

Primary and secondary are roles, not automatic quality rankings. The production needs provenance and interpretation for both.

Evidence, visual reference, and rights are separate

One archival photograph may serve three different production jobs.

It may support a claim about a building. It may guide clothing and lighting in a reconstruction. It may be licensed for direct use in the film.

Those conclusions must be recorded separately. A factual object is not automatically reusable. A public website is not automatically public domain. A visual mood reference is not evidence.

The source-pack guide recommends labeling each item by evidence role, visual role, and rights status so one valid conclusion does not silently create another.

Review should happen before narrative commitment

The pack should be reviewed before the story blueprint and narration become expensive commitments.

This is where the team can discover that the hook depends on weak evidence, two important sources conflict, a central quotation has no reliable origin, or the available visual record cannot support the planned reconstruction.

The correct response may be to narrow the thesis, attribute disagreement, change the opening, or choose another story. Generating more scenes is not a research solution.

The pack remains alive through publication

New evidence can arrive during scripting or review. When it does, record the addition, reviewer, affected claims, and production artifacts that changed.

At final approval, archive the accepted pack with the claim ledger, script, narration, rights record, disclosure decision, correction log, and exact export. That chain makes post-publication corrections possible without pretending that the production is infallible.

Onira currently remains research-assisted; creator verification is required. The existence of an internal research stage does not yet create a user-facing source receipt for every claim.

Creators can still build a rigorous external process using the documentary research brief and the fact-checking guide.

The objective is not a longer bibliography. It is a smaller, stronger evidence boundary that survives contact with the final film.