Product Design Specification¶
Document type: Product design specification (target state). Not a shipped feature list. See ROADMAP_MAPPING.md for release mapping and SHIPPED.md for what works today.
1. Product Vision¶
OntoCode is a semantic engineering environment for ontology and knowledge graph development. It should feel like a modern IDE, not a legacy ontology editor.
The long-term vision is to create the JetBrains/Figma/Cursor equivalent for semantic engineering:
- JetBrains-level navigation, refactoring, and diagnostics.
- Figma-level graph/canvas interaction.
- Cursor-level AI assistance.
- GitHub-level semantic collaboration.
- DataGrip-level querying.
- Rust-level performance.
2. Product Principles¶
2.1 Context Over Windows¶
Every view derives from the current semantic context. Selecting an entity updates the editor, inspector, graph, reasoning, documentation, references, and AI suggestions.
2.2 Workflows Over Panels¶
Users do not think in "class tree", "graph", or "reasoner" panels. They think in tasks:
- Understand this entity.
- Fix this diagnostic.
- Refactor this model.
- Review this change.
- Explain this inference.
- Publish this documentation.
2.3 Workspace Over Files¶
OntoCode treats an ontology repository as a semantic workspace. Files are implementation details. Users navigate entities, relationships, queries, diagnostics, modules, and documentation.
2.4 AI as Collaborator¶
AI is embedded in every workflow. It explains, reviews, repairs, documents, refactors, and teaches, but never applies ontology changes without preview and approval.
2.5 Safe Transformation¶
Every significant change is previewable, undoable, and reasoning-aware.
3. User Personas¶
3.1 Ontology Engineer¶
Needs efficient editing, reasoning, refactoring, navigation, and review.
3.2 Domain Expert¶
Needs simplified semantic views, documentation, diagrams, comments, and guided workflows.
3.3 Data Engineer¶
Needs SQL/SPARQL querying, exports, validation, and integration with pipelines.
3.4 Researcher¶
Needs graph exploration, explanations, documentation, and provenance.
3.5 Platform Developer¶
Needs plugin APIs, SDKs, diagnostics, tests, and stable contracts.
4. Core Workspaces¶
4.1 Entity Workspace¶
The main semantic object editor. Shows overview, hierarchy, relationships, constraints, annotations, documentation, history, references, reasoning, and AI.
4.2 Graph Workspace¶
Interactive semantic canvas. Supports persistent layouts, semantic overlays, reasoning overlays, grouping, AI exploration, saved views, and presentation mode.
4.3 Query Workspace¶
DataGrip-like environment for OntoSQL, SPARQL, SHACL, and future query languages.
4.4 Reasoning Workspace¶
Compiler-like health dashboard with build pipeline, diagnostics, explanations, quick fixes, and reasoning history.
4.5 Review Workspace¶
Semantic pull requests, semantic diffs, review threads, approvals, merge checks, and AI review.
4.6 Documentation Workspace¶
Author, preview, validate, publish, and generate documentation.
4.7 AI Workspace¶
Long-running AI workflows for project-wide documentation, review, refactoring, onboarding, and architecture analysis.
5. Global Application Shell¶
+--------------------------------------------------------------------------------+
| Menu | Search / Command Palette | Workspace Switcher | AI | Git | User |
+--------------------------------------------------------------------------------+
| Explorer | Active Workspace | Inspector |
| | | |
| | | |
+--------------------------------------------------------------------------------+
| Problems | Query | Graph | AI | Git | Output | Terminal | Notifications |
+--------------------------------------------------------------------------------+
6. Navigation Model¶
- Universal search is the primary navigation surface.
- Breadcrumbs provide location awareness.
- Back/forward history is semantic, not file-based.
- Favorites and recent entities support deep ontology workflows.
- Jump-to-definition and find-references work across ontology formats.
7. Quality Bar¶
The application should feel:
- Fast
- Calm
- Professional
- Discoverable
- Accessible
- AI-native
- Safe for large-scale modeling
8. Non-Goals¶
- Do not clone Protégé UI patterns blindly.
- Do not make every feature a separate panel.
- Do not require users to understand serialization formats before they can be productive.
- Do not let AI apply hidden changes.
- Do not expose raw parser or reasoner errors without interpretation.