OpenSPAD Documentation
Chapter 02 · Semantic Core · Specification Draft 0.1

Concept Governance
& Versioning

Define, compose and maintain semantically meaningful Concepts across data sources, projects and future federated servers — with traceable revisions and controlled publication.

Concept GraphVersioned RevisionsBusiness RulesSource AlignmentGovernance
This document describes the OpenSPAD 7.8.53 pilot separately from planned capabilities. It is not a production implementation claim.
01 · Foundation

What is a Concept?

A Concept is a reusable, identifiable, semantically enriched building block from which OpenSPAD assembles information requirements and Use Cases. It is not merely a label, an IFC class, an OWL class or a source-file row.

Semantic meaning

Identity, label, definition, domain, classification and relationships to source ontologies.

Information structure

Classes, property sets, properties and possible constraints, linked to provenance.

Governed reuse

Revision, approval state, combination rules and stable references from Use Cases.

A project-specific imported IFC model or Excel dataset is a source resource; it is not automatically a globally published Concept. Globally reusable Concepts require administrative governance.
02 · Information model

A Concept is a governed subgraph

The canonical property graph represents Concept identity separately from its revisions and source-specific evidence. This diagram is the target conceptual model; individual relationships may not yet exist in the stored 7.8.53 schema.

Concept IdentityStable ID · NamespaceRevisionDraft / PublishedSource EvidenceOWL · bSDD · IFC · ExcelGraph RulesChild / ExclusionSelected IFC ClassesProperty sets / PropertiesConsuming Use CasesRevision-pinned referencesValidation & Approval Evidence

Identity vs. revision

A stable Concept identity can have multiple revisions. A Use Case should retain a reference to the revision it was created with until an explicit migration is approved.

Relationship semantics

Do not confuse class membership, source alignment, Concept hierarchy and a workflow dependency between two Use Cases. Each is a distinct edge type with distinct governance.

03 · Controlled change

Concept lifecycle and safe revisions

A new Concept is reviewed before publication. A correction produces a new revision, preserving existing published references and allowing administrators to assess the impact.

01 · CREATENew draft identity or revision
→
02 · ENRICHSources, classes, properties
→
03 · CHECKRules and affected Use Cases
→
04 · REVIEWAdministrator acceptance
→
05 · PUBLISHImmutable version for reuse
Example: Heating Systems revision 1 contains a direct ALLOWS_CHILD → Boilers relationship in the 7.8.53 pilot. Changing allowed children should create and evaluate a new draft revision; already saved Use Cases must not be silently rewritten.
04 · Graph composition

Rules determine what may connect

7.8.53 limited pilot

Direct edge evaluation

The persisted graph pilot evaluates direct ALLOWS_CHILD / EXCLUDES rules from selected revisions. A permitted child in the draft pilot is not automatically production-approved.

  • Allowed: Heating Systems → Boilers
  • Blocked: Heating Systems → Accessibility (not whitelisted)
  • Root anchor eligibility is not yet validated.
Target architecture

Complete composition engine

Broaden rule evaluation to include root eligibility, Concept types, multi-hop dependencies, REQUIRES, PRECEDES, global exclusions, cycle detection and explainable failure traces.

Use Case workflow dependencies must remain separate from Concept-child rules.

Rule evaluation result should explain why

InputEvaluationOutput
Parent revision + candidate childSelected revision, publishing mode, direct rule setAllowed / blocked, reason, responsible rule, provenance
Composite graph (future)Root, type, dependency and global checksExplainable validation findings and suggested repair
05 · Evolvability

When a new data source arrives

External inputs produce normalized graph fragments and alignment candidates. They must not overwrite published Concept revisions or bypass governance.

INGESTIFC / OWL / bSDD / Excel / IDS / API
→
NORMALIZECanonical graph fragment + origin
→
ALIGNFind existing Concepts / conflicts
→
PROPOSENew evidence or draft revision
→
APPROVEAdministrator-controlled publish

Global resources

General OWL ontologies, IFC schema definitions, classifications and published reference Concepts can be reused across projects.

Project resources

An uploaded IFC model, Excel sheet or project ontology contributes to a project-specific graph. This does not automatically change global knowledge.

An IFC or Excel file is a data source, not automatically an ontology. OpenSPAD interprets it through mappings and records its resulting semantic representation with provenance.
06 · Administrator manual

Two core governance workflows

Create a Concept

  1. Open the Concept library and establish a unique identity.
  2. Specify meaning, type and applicable domain.
  3. Attach approved data sources and select classes / properties.
  4. Define parent-child rules and exclusions.
  5. Inspect the proposed graph and diagnostics.
  6. Publish only when required checks and approval exist.
Full workflow specification

Correct an existing Concept

  1. Select the exact existing Concept and revision.
  2. Create a new draft revision, keeping published history.
  3. Change sources, mappings, constraints or rules.
  4. Compare differences and assess impacted Use Cases.
  5. Approve, publish and notify dependent projects.
  6. Preserve old references until explicit migration.
Full workflow specification

Role boundary

ActionAdministratorBIM ManagerStandard User
Create / correct / publish global ConceptsYesNoNo
Select approved Concepts in project Use CasesYesYesVia references
Create / edit project Use CasesYesYesNo
Assemble approved Use Cases into IDMYesYesYes
Manage source registration / backupYesNoNo
07 · Long-term architecture

Federated Concepts without losing ownership

A manufacturer or partner may manage its own graph, expose only selected resource fragments and maintain its own revisions. Consumer servers link to authorized remote resources rather than assuming complete ownership of the provider’s data.

Stable identities

Global resource identifiers distinguish an entity from an aligned, equivalent or related entity on another server.

Provider governance

Each source owner controls published versions, availability and access policy for its resources.

Future security

Tenant isolation, resource-level permissions, authorized query federation and policy-aware caching are planned, not implemented in the 7.8.53 pilot.

08 · Evidence and maturity

What is available today — and what is not

CapabilityAssessmentEvidence / qualification
Concept Editor and revision-related routesCode/routes observedRegistered in Administration; individual user flows require separate testing.
Persisted Rule Reader and Graph Pilot 7.8.52/53Pilot verifiedDirect draft-mode rule evaluation and observed successful/blocked examples.
New-source alignment and global Concept approval workflowPartly designedFull end-to-end governance and change propagation not verified.
Cross-server Concept federation and multi-tenant policiesFutureTarget architecture only.
Use Case revision migration and full rule enginePlannedRoot, multi-hop and global rule checks remain open.
Documentation principle: every feature must state whether it is verified, a pilot, planned or future architecture. Architecture diagrams illustrate design intent, not deployment status.