GRYPS · CHANGELOG · GRYPS Methodology v0.6
Methodology changelog
How the GRYPS research model and documentation evolve. Current methodology framework is v0.6; the deterministic scoring engine remains reproducible under its own version tag.
GRYPS is an experimental research prototype. Results are indicative and should not be interpreted as commercial procurement advice, coverage certification, or engineering advice.
Current
GRYPS Methodology v0.6
Scoring engine
deterministic · v0.4
Signature model tag: gryps-signature-v0.4
v0.6
2026-10-13CurrentTwo-layer Signature (Model v0.4)
Scoring engine: deterministic-v0.4
- Split score into Site Conditions (fixed) and Mitigation Quality (actionable); composite always shown with both layers.
- Reweight mitigation factors (35 / 25 / 20) via ratio remap of v0.3 inputs; grade bands A≥85 · B 70–84 · C 55–69 · D 40–54 · F <40.
- Confidence ranges from evidence band; Layer-B levers; adjacent evidence panel (EU-DEM excluded from 0–100).
- AI-assisted interpretation becomes opt-in; Art. 50 disclosure moved off the score badge.
v0.5
2026-09-11Portfolio / Demonstration Layer
Scoring engine: deterministic-v0.3
- Portfolio case study: problem, research question, approach, live prototype links, learnings, and explicit non-production limits.
- Demonstration framing for AI / digital solutions / automation / business-analyst roles — not monetization.
- Public Research & Prototype posture page; commercial features remain outside current scope. Detailed project-control notes stay out of the product UI.
v0.4
2026-09-11Prototype Research Quality
Scoring engine: deterministic-v0.3
- Formal research documentation set: methodology, data sources, assumptions, limitations, changelog.
- Data provenance records (source, date accessed, type, how GRYPS uses it).
- Explicit prototype disclaimer: indicative research — not procurement, certification, or engineering advice.
- Research Workspace, mission scenarios, and evidence chain remain non-commercial analytical surfaces.
v0.3
2026-Q1–Q3Deterministic Signature + Advisor Intelligence
Scoring engine: deterministic-v0.3
- Scoring engine deterministic-v0.3 / gryps-signature-v0.3: redundancy, latitude, operational profile, provider confidence, hard caps.
- Advisor Intelligence: recommendation package, score explanations, mission priorities, provider comparison.
- Evidence chain: environment → research → characteristics → scoring → recommendation.
- Research Library, Connectivity Intelligence map, mission scenarios, local Research Workspace.
v0.2
2025–2026Assessment-first product language
- Canonical CTA: Generate Resilience Signature; confidence ≠ availability; claim hygiene.
- MapLibre ops console, Providers catalog, knowledge notes, Advisor funnel (abbreviated → unlock).
v0.1
2025Initial research prototype
- First Nordic/Arctic connectivity resilience scoring explorations and non-commercial R&D posture.