# ContentDNA Provenance Specification v1.0

ContentDividend application release: 15.78.0 — Phase F Step 6

## Purpose

ContentDNA Provenance defines an open, portable envelope for identifying a
publisher-registered content record and binding that record to reproducible
fingerprints and provenance metadata.

The format is open. ContentDividend's proprietary similarity matching,
detection intelligence, corpus comparisons, ranking, and commercial evidence
systems are not part of this specification.

## Core identifiers

- `schema_version`: `contentdividend.contentdna-provenance.v1`
- `fingerprint_profile`: `contentdividend.canonical-text.v1`
- `fingerprint_algorithm`: `sha-256`
- `proof_suite`: `eddsa-ed25519-contentdividend-2026`

## Portable canonical-text profile

The open portable fingerprint is computed over publisher-submitted text after:

1. Unicode NFKC normalization.
2. CRLF/CR converted to LF.
3. NUL characters removed.
4. Horizontal whitespace collapsed to one ASCII space.
5. Leading/trailing horizontal whitespace removed per line.
6. Runs of three or more newlines collapsed to two.
7. Entire value trimmed.

The resulting UTF-8 bytes are hashed using SHA-256.

This portable digest is separate from any internal ContentDividend similarity
fingerprint or proprietary matching representation.

## Provenance semantics

A record can identify:

- publisher
- registered site
- canonical source URL
- registry entry
- registration/observation time
- content type/title
- portable SHA-256 fingerprint
- existing registry SHA-256 binding
- Publisher Rights Passport
- URVP rights-verification endpoint
- Ed25519 proof

## Rights boundary

A provenance record proves neither ownership as a matter of law nor permission
to use content. It does not grant a license. Current ContentDividend rights,
agreements, and exact publisher authorization remain authoritative.

## External provenance standards

ContentDNA Provenance is designed to coexist with provenance ecosystems such as
C2PA Content Credentials and W3C PROV. ContentDividend does not claim C2PA
conformance/certification or W3C endorsement in F6.

## Open-source boundary

Open:
- provenance record schema
- canonical-text algorithm
- portable SHA-256 calculation
- local proof verification helpers

Not open:
- ContentDividend proprietary similarity algorithms
- internal match ranking
- private corpus comparisons
- crawler intelligence
- private registry/buyer data