Digital Preservation Domain Ontology

FLOW ALGEBRA
TITLE: Digital Preservation Domain Ontology
VERSION: 1.0
DOMAIN: Information Domains / Digital Preservation
SOURCE_URL: https://kirethresearch.org/booksai-core/ontologies/domain-ontologies/information-domains/digital-preservation-domain-ontology.html
STATUS: Domain-based draft; source page content not independently verified.
AUTHORS: Mark Higgins — Sky (Recursive AI)
PROJECT: Kireth Semantic Library

PURPOSE
Model the long-term stewardship, accessibility, authenticity, integrity, usability and provenance of digital information objects across technological and organisational change.

TYPED ENTITIES
DigitalObject: entity {identifier: URI, title: string, created_at: datetime?, media_type: string?}
Representation: entity {identifier: URI, format: string, version: string?}
Bitstream: entity {identifier: URI, size_bytes: integer?, checksum: string?}
PreservationRepository: entity {identifier: URI, name: string}
PreservationPackage: entity {identifier: URI, package_type: enum[SIP,AIP,DIP]}
PreservationEvent: entity {identifier: URI, event_type: string, occurred_at: datetime}
PreservationAgent: entity {identifier: URI, role: string}
PreservationPolicy: entity {identifier: URI, effective_from: datetime?}
FormatRegistryEntry: entity {format_id: string, risk_level: string?}
FixityCheck: entity {identifier: URI, algorithm: string, expected_digest: string, observed_digest: string, checked_at: datetime}
Migration: entity {identifier: URI, source_format: string, target_format: string, performed_at: datetime}
AccessRequest: entity {identifier: URI, requested_at: datetime, authorised: boolean?}
RightsStatement: entity {identifier: URI, rights_holder: string?, restrictions: string?}
ProvenanceRecord: entity {identifier: URI, recorded_at: datetime, source: URI?}

TYPED RELATIONSHIPS
has_representation(DigitalObject, Representation)
contains_bitstream(Representation, Bitstream)
stored_in(PreservationPackage, PreservationRepository)
packages_object(PreservationPackage, DigitalObject)
subject_of_event(DigitalObject, PreservationEvent)
performed_by(PreservationEvent, PreservationAgent)
governed_by(PreservationRepository, PreservationPolicy)
uses_format(Representation, FormatRegistryEntry)
checks_bitstream(FixityCheck, Bitstream)
migrates_from(Migration, Representation)
migrates_to(Migration, Representation)
subject_to_rights(DigitalObject, RightsStatement)
documented_by(DigitalObject, ProvenanceRecord)
requests_access_to(AccessRequest, DigitalObject)

CONSTRAINTS
C1: DigitalObject.identifier and PreservationEvent.identifier MUST be unique within the repository scope.
C2: A successful fixity result requires a supported digest algorithm and exact expected/observed digest equality.
C3: Migration MUST retain the source representation or an auditable reference to it and record the transformation event.
C4: Access is permitted only when applicable rights, policy and authorisation requirements are satisfied.
C5: Preservation metadata MUST distinguish observed evidence from inferred conclusions.
C6: AIP, SIP and DIP roles MUST remain distinct even if one implementation reuses physical storage.

INFERENCE RULES
R1: IF FixityCheck.expected_digest == FixityCheck.observed_digest AND algorithm_valid THEN infer fixity_verified_at(Bitstream, checked_at); do not infer authenticity solely from fixity.
R2: IF digest mismatch THEN infer integrity_alert(Bitstream) AND require investigation; do not infer malicious modification.
R3: IF Representation format is unsupported or obsolete THEN infer preservation_review_required(Representation); migration is one possible response, not an automatic obligation.
R4: IF Migration is validated AND documented THEN infer successor_representation(target, source) with preservation-event provenance.
R5: IF authorised request AND usable dissemination representation exists AND rights permit THEN infer access_eligible(DigitalObject).
R6: IF provenance chain has missing links THEN infer provenance_incomplete(DigitalObject), not provenance_false(DigitalObject).
R7: IF redundant independent copies exist AND recent fixity checks pass THEN infer reduced_bit_loss_exposure, conditional on failure independence.

CONFIDENCE
confidence: float in [0,1]
confidence_basis: enum[direct_observation, validated_metadata, expert_assessment, model_inference]
RULE: Confidence represents evidential support, not a guarantee of future preservation or legal authenticity.

SEMANTIC_EQUIVALENCE
DigitalObject ~= DigitalResource WHEN identity criteria match.
PreservationCopy != BackupCopy by default: purpose and retention obligations differ.
Authenticity != Integrity: an unchanged object may still lack verified origin.
Migration != Emulation: migration transforms representation; emulation recreates an execution environment.
AIP != DIP: archival and dissemination packages serve different roles.

PROVENANCE
For each asserted fact record: source_identifier, agent, event_timestamp, method, evidence_reference, version and confidence_basis.
Preserve immutable audit records where feasible; corrections append superseding events rather than silently rewriting history.

FLOW_HASH
algorithm: SHA-256
input: canonical UTF-8 serialisation of versioned ontology definitions and rules, excluding the stored hash value.
output: lowercase hexadecimal digest
NOTE: Hash must be computed by the consuming system; no fabricated digest is asserted here.

UPDATE_MODEL
strategy: append-only provenance + versioned ontology releases
on_change: validate schema -> record change event -> recompute FLOW_HASH -> update dependency mappings -> publish version
on_conflict: retain conflicting claims with source, timestamps and resolution status until reviewed

AI_CONSUMPTION
parse_mode: typed_entities_and_relations
reasoning: evidence-grounded, rule-bounded, provenance-aware
retrieval_keys: [identifier, format, repository, event_type, date, rights_status]
output_requirements: [claim, supporting_evidence, confidence, provenance, uncertainty]
safety: Never infer preservation success, legal admissibility, or authenticity from a checksum alone.

END FLOW ALGEBRA