Edge-canonical conversational knowledge-building platform
graph TD P0["Phase 0 — Scaffolding ✓"] --> P05["Phase 0.5 — Metadata Types ✓"] P0 --> P1["Phase 1 — Identity Simplification ✓"] P1 --> P2["Phase 2 — NLParser + Classifier ✓"] P3 --> P7["Phase 7 — DescriptionEngine ✓"] P2 --> P10["Phase 10 — ExportEngine (A) ✓"] P10 --> P10b["Phase 10b — ERS Core (A/B) ✓"] P1 --> P3["Phase 3 — StateAdapter (B) ✓"] P3 --> P4["Phase 4 — Validator (B) ✓"] P4 --> P4b["Phase 4b — OCE/IEE Stubs (B) ✓"] P4b --> P5["Phase 5 — Classification (B) ✓"] P5 --> P6["Phase 6 — Properties (B) ✓"] P5 --> P9["Phase 9 — Relationships (B) ✓"] P3 --> P11["Phase 11 — Sessions (C)"] P11 --> P12["Phase 12 — Federation (C)"] P2 --> P8["Phase 8 — Integration ✓ ★"] P5 --> P8 P7 --> P8 P1 --> IVNE["IVNE — OWL Compiler ✓"] P10b --> IVNE P8 --> P13["Phase 13 — M2M Protocol"] P13 --> P14["Phase 14 — Ecosystem"] style P0 fill:#1a3a2a,stroke:#3dd68c,color:#3dd68c style P05 fill:#1a3a2a,stroke:#3dd68c,color:#3dd68c style P1 fill:#1a3a2a,stroke:#3dd68c,color:#3dd68c style P2 fill:#1a3a2a,stroke:#3dd68c,color:#3dd68c style P3 fill:#1a3a2a,stroke:#3dd68c,color:#3dd68c style P4 fill:#1a3a2a,stroke:#3dd68c,color:#3dd68c style P4b fill:#1a3a2a,stroke:#3dd68c,color:#3dd68c style P5 fill:#1a3a2a,stroke:#3dd68c,color:#3dd68c style P6 fill:#1a3a2a,stroke:#3dd68c,color:#3dd68c style P7 fill:#1a3a2a,stroke:#3dd68c,color:#3dd68c style P8 fill:#1a3a2a,stroke:#3dd68c,color:#3dd68c style P9 fill:#1a3a2a,stroke:#3dd68c,color:#3dd68c style P10 fill:#1a3a2a,stroke:#3dd68c,color:#3dd68c style P10b fill:#1a3a2a,stroke:#3dd68c,color:#3dd68c style IVNE fill:#1a3a2a,stroke:#3dd68c,color:#3dd68c
Type any term to see the 7-step normalization pipeline in real time. Each step shows the intermediate result — changes are highlighted in green.
| Input | Expected | Note |
|---|
Select a type factory, fill in parameters, and generate a JSON-LD node.
All computation runs in your browser — the bundle is loaded from dist/fandaws.js.
Phase 6 property attachment with scope narrowing. Type a "has" statement, then answer scope narrowing prompts to see where the property attaches.
Phase 7 auto-generates human-readable descriptions from graph structure. Select a concept, add properties, and see the description update live. Supports standard, root, and process templates with automatic article selection ("a" vs "an").
Add a relationship to trigger the process template instead of standard.
Phase 9 custom relationship pipeline. Type a relationship statement like "Dogs chase cats" to see verb normalization, sub-relationship detection, and graph mutation.
Phase 10 ExportEngine. Export a knowledge graph to standard ontology formats: SKOS, OWL, RDF/XML, or Turtle. The graph below is preloaded with a 5-concept animal taxonomy including properties and a relationship.
Select a format and click Export.
The Epistemic Register Service (ERS) solves the Normative-Axiomatic Conflation problem by routing every restriction node to one of three registers: R1 Axiomatic (definitional — exceptions are contradictions), R2 Normative (statistically typical — exceptions are expected), or R3 Aspirational (value judgments — never auto-assigned). The 6-step pipeline uses BFO alignment, session domain, and teleological detection — no LLMs.
Test user-initiated register overrides. R3 requires a worldview context tag.
The Ingestion, Validation & Normalization Engine compiles OWL 2 DL ontologies into native Fandaws concepts. Select a scenario to see the three-phase transformation pipeline: Flatten (τF) → Lift Restrictions (τR) → Normalize (τN). All computation is deterministic — same input always produces byte-identical output.
Phase 11 implements session state machine management: pause/resume, nested negotiation, abandon with cascade, concurrent session limits, and expiration. All lifecycle logic is pure functions — the OrchestrationAdapter coordinates persistence via the StateAdapter. Run the scenarios below to observe state transitions and dialogue history accumulation.
Full utterance-to-graph pipeline via SynchronousOrchestrationAdapter. Type natural language statements to build a knowledge graph interactively. The adapter routes each utterance to the correct pipeline (classification, property, or relationship).
| Suite | Passed | Failed | Duration |
|---|