Skip to content
🚧 Early alpha — building the foundation. See the roadmap →

Deep research synthesis — ontology meta-system, plugin state, E2E testing

Created Updated

Research 1: Ontology lifecycle management — does anything solve this?

Section titled “Research 1: Ontology lifecycle management — does anything solve this?”

Finding: The meta-problem is NOT solved by any existing tool.

Individual pieces exist:

  • MITRE ATT&CK has machine-readable deprecation chains (STIX 2.1) — gold standard
  • OSCAL uses SemVer with content validity guarantees
  • SCF STRM uses interlingua/pivot approach for 175+ frameworks
  • CISO Assistant has YAML library versioning

But NO tool:

  • Classifies HOW an ontology evolves
  • Auto-selects a handling strategy based on that classification
  • Provides a registry/marketplace for community framework configs
  • Implements automated migration between versions

What we need to build from scratch:

  1. EvolutionPattern taxonomy (release cadence, breaking change policy, ID stability, changelog format)
  2. Evolution profile registry (YAML/JSON per framework)
  3. Migration strategy engine (given old + new version → suggest handling)
  4. Community config marketplace

Key insight: SCD types from data warehousing are more mature than any ontology-specific approach. Apply them directly.

Solid MVP (0.1.0) but significant gaps:

ComponentCompletenessNotes
Generation engine95%Production-ready, _crosswalker metadata, YAML serialization
Config manager100%Save/load/fingerprint/match, schema migration v0→v1
CSV parser100%PapaParse with streaming, column analysis
Settings schema100%50+ settings with defaults
Debug logging100%File-based, toggleable
Import wizard UI~30%Steps 2-4 rendering incomplete
Settings tab UI~30%Missing 70% of sections
Config browser modal~30%Only header logic visible
Transform system0%Schema defined (20+ types), zero implementation
XLSX parser0%Package installed, no integration
JSON parser0%Not started

19 unit tests, 2 E2E smoke tests. Harness ready but no workflow coverage.

Research 3: E2E testing with wdio-obsidian

Section titled “Research 3: E2E testing with wdio-obsidian”

Current state: Config exists, smoke tests exist, but not ready for iterative dev.

What wdio-obsidian can test:

  • Plugin loading, command registration
  • Modal interactions (open, fill forms, click buttons)
  • File operations through Obsidian API
  • Command execution and UI state verification
  • Vault reset between tests

What’s missing for Crosswalker:

  • Test helpers (vault reset, file upload, modal interaction utilities)
  • Import wizard E2E tests (the main workflow)
  • Test data fixtures (small/medium/large CSVs, multi-sheet XLSX)
  • TypeScript config for test files
  • GitHub Actions E2E workflow

Estimated setup: 3-4 hours for minimum viable E2E, 5-7 hours for CI-ready.

  1. Architecture decisions FIRST — the ontology evolution classification and config schema must be designed before building more features
  2. Complete the UI before adding features — wizard, settings tab, and config browser are 30% done
  3. E2E testing enables iteration — set it up early, write tests alongside features
  4. Transform system is Phase 2 — schema is solid, implementation can follow
  5. The “meta” layer is novel — no existing tool to map to, must build from first principles