1. Purpose
Name the practical question, affected people, intended benefit, and excluded decisions.
Authoring a Semantic Instrument starts with a human purpose — a clinical checklist, a community referral rule, a professional rubric, or a multi-party agreement — and helps contributors make vocabulary, evidence, authority, privacy, and review explicit before software executes a single step.
This guide moves from a plain-language overview of the authoring workflow to the Four-Plane Identifier Fabric (IDF-01–IDF-05), developer morphisms, and formal SHACL and compiler firewall specifications.
Authoring path
The workspace people need
Purpose, affected people, benefit, risks, excluded outcomes, and the narrow question the tool is authorised to answer. The executable rule stays traceable to this readable brief.
A place for people to contribute, qualify, dissent, authorise, delegate narrowly, challenge, or withdraw. Legal entities and AI tools always preserve traceable human principal accountability.
A visual, inspectable connection between terms, ontologies, claims, sources, and rules. Domain experts can start visually while ontology engineers work directly with RDF, N3, and SHACL.
Shows what was known at decision time ($T_0$), what changed later ($T_1$), which assumptions expired, and which material changes trigger review or remedy.
Specifies what a routine verifier learns, what remains in Sanctuary custody, who may request controlled disclosure, under which procedure, and with what auditable receipt.
Tests examples, counter-examples, missing inputs, conflicts, and temporal transitions before activation, showing clearly why a run is ready, held, or escalated.
Layer 2 · Architecture · Identifier Fabric (IDF-01–IDF-05)
A common failure in software systems is collapsing a person into an email address, a cryptographic key, a wallet, or a probabilistic score. QualiaDB enforces a strict Four-Plane Identifier Fabric (ontologies/identifier_fabric_shapes.ttl) so authoring tools model human agency, evidentiary claims, network locators, and cryptographic instruments on disjoint planes connected only by typed evidence morphisms.
NaturalAgent · LegalEntity · SystemAgent
Represents accountable actors on the Who plane. idf:NaturalAgent is framed as "living-SHACL" and is structurally disjoint from owl:Thing, idf:Instrument, and idf:SpatiotemporalHandle.
Invariant: Never merged by similarity, handle, wallet, location, or biometrics. owl:sameAs is forbidden on the Who plane (SameAsError::WhoPlaneForbidden).
VerifiableCredential · ClaimOpinion · RoleCapacityGrant
Holds scoped, time-bounded assertions (idf:AssertionClaim, idf:AccountabilityArtifact, idf:RelianceRecord, idf:AssumptionChain, deontic and epistemic norms) with explicit issuer provenance (idf:assertedBy) and subject reference (idf:aboutSubject).
Invariant: Verifiable Credentials prove cryptographic origin integrity of an issuer's statement, not universal ontological truth.
SpatiotemporalHandle · DNI · did:qi Endpoints
Covers rotatable, relationship-scoped network and directory locators (idf:DigitalNetworkIdentifier, idf:RelationScopedLocator, idf:EnvironmentPredicate, DNS, email, and did:qi endpoints) anchored by idf:frameId and idf:epochTick.
Invariant: Revoking or rotating a handle closes the communication channel without erasing or invalidating the underlying Entity.
Instrument · AiAgent · MachineDevice · Keys
Cryptographic keys, hardware tokens, wallets, cancelable biometric templates, and software/AI tools classified by InstrumentProofClass (CryptographicOriginIntegrity, HardwareChannelPossession, SemanticClaimEvaluation) and closed KeyRole (IdentityAssertion, CredentialSigning, TransportHandshake, StorageEncryption, RecoveryBootstrap).
Invariant: Every idf:AiAgent resides on the Instrument plane and requires an explicit idf:sponsorPrincipal.
Layer 3 · Typed Cross-Plane Evidence Morphisms
Rather than equating a person with a key or a network handle, statements traverse planes only through typed evidence predicates:
@prefix idf: <urn:qualia:idf:> .
# Claim Plane → Instrument Plane (cryptographic origin proof)
ex:claim-042 a idf:ClaimOpinion ;
idf:plane "Claim" ;
idf:assertedBy ex:issuer-college ;
idf:aboutSubject ex:clinician-entity ;
idf:assertedAbout ex:clinician-entity ;
idf:supportedByInstrument ex:ml-dsa-65-signing-key ;
idf:observedAtHandle ex:dni-clinic-session-09 .
# Attempting owl:sameAs on the Who plane returns:
# Err(SameAsError::WhoPlaneForbidden)
Layer 4 · Formal SHACL & F5 / F6 Anti-Feature Firewalls
Both the F5 VibeScript compile-time verifier and the F6 runtime ingress/query firewall reject surveillance-scoring and entity-collapsing predicates (person_score, identity_confidence, same_person):
# From ontologies/identifier_fabric_shapes.ttl (Packet IDF-01)
idf:NaturalAgentShape a sh:NodeShape ;
sh:targetClass idf:NaturalAgent ;
sh:property idf:NaturalAgent-planeProp ; # sh:hasValue "NaturalAgent"
sh:property idf:NaturalAgent-framingProp ; # sh:hasValue "living-SHACL"
sh:not idf:NaturalAgent-notOwlThing ; # disjoint from owl:Thing
sh:not idf:NaturalAgent-notInstrument ; # disjoint from idf:Instrument
sh:not idf:NaturalAgent-notHandle . # disjoint from SpatiotemporalHandle
Relation to ontology editors
Specialist environments such as Protégé excel at formal class hierarchies and axiom engineering. QualiaDB's human-centric authoring workspace retains full RDF, N3, and SHACL interoperability while surrounding the formal model with purpose, party perspectives, Identifier Fabric plane separation, disclosure rules, and verifiable run receipts.
| Concern | Ontology-centred workbench | Human-centric instrument environment |
|---|---|---|
| Primary starting point | Classes, properties, axioms, imports, and reasoners. | A human purpose, affected parties, intended use, and prohibited outcomes—then the meanings and rules required. |
| Identifier separation | Often uses uniform owl:Thing individuals and owl:sameAs links. | Enforces the Four-Plane Identifier Fabric (Entity, Claim, Handle, Instrument) and blocks Who-plane merging. |
| Authority and agreement | Usually handled outside the model editor. | First-class authorship records: accountable human roles, mandates, party positions, ratification, dissent, and review. |
| Privacy and disclosure | Delegated to external deployment infrastructure. | Models Sanctuary custody (0x0040), ZK predicates, purpose-scoped release conditions, and receipts beside the rule. |
| Operational use | Model validation and inference are central. | Includes preflight, bounded execution, held-run safeguards, human escalation, temporal review, and cryptographic receipts. |
Interaction principles
Delivery architecture
The authoring experience combines three focused layers: the native Desktop host that manages local hardware and Sanctuary custody boundaries, the POET visual relational workspace where collaborators map concepts and agreements, and the QualiaDB engine that validates SHACL shapes, evaluates modal logic, and issues receipts.
LOCAL DESKTOP HOST
Launches POET, manages local .q42 v4 volumes, key vaults, renderer capabilities, and model runtimes, and enforces FLAG_SANCTUARY publication locks.
POET WORKSPACE
Presents a visual manifold where concepts, sources, claims, rules, party positions, and consequences connect through typed semantic and modal relations.
QUALIADB / WEBIZEN ENGINE
Compiles SHACL and N3 rules, enforces the Four-Plane Identifier Fabric and F5/F6 firewalls, evaluates 21 logic modalities, and generates verifiable receipts.
An author may invite a local or permitted generative agent (idf:AiAgent on the Instrument plane) to suggest terms, draft plain-language summaries, propose ontology alignments (cml:Proposed), or generate edge-case tests inside a bounded drafting lane.
The author selects which ontologies, records, references, and tool permissions the agent may access.
Every proposed node, connector, rule, or mapping arrives as an attributed draft that people can accept, amend, or reject.
Only accountable human principals can ratify an agreement, attest a concept (cml:Attested), authorise disclosure, or activate an instrument.
WASM POET · Collaborative Manifolds
POET's WebAssembly build lets collaborators inspect and co-author a selected semantic manifold directly in the browser before exporting or activating the package in a local Desktop environment.
Participants inspect the same scoped graph and propose nodes, relations, comments, alternatives, and tests. Competing interpretations remain attributable and versioned until resolved.
Connectors carry formal semantics—“is defined by”, “supports”, “defeats”, “is permitted by”, or cross-plane evidence morphisms (idf:supportedByInstrument, idf:assertedAbout).
Browser sessions operate on declared ontology packs and sample or permitted records, while custodial keys and Sanctuary volumes remain in the authorised local host.
Continue
The Semantic Instruments guide details the runnable package lifecycle, while the Agreement and Values Credentials pages show how multi-party consent and human-rights norms compile into verifiable .q42 v4 volumes.