Skip to content

Releases: ericsocrat/decision-os

v0.1.0 — Initial Public Release

25 Feb 15:37
@ericsocrat ericsocrat
v0.1.0
21daa07

Choose a tag to compare

View all tags
v0.1.0 — Initial Public Release Latest
Latest

Decision OS v0.1.0 — Initial Public Release

Decision OS is a structured, deterministic decision-making tool that eliminates gut-feel from important choices. Built with Next.js, TypeScript, and a mathematically rigorous scoring engine.

Core Engine

  • Weighted Sum Model (WSM): Deterministic scoring — same inputs always produce the same output
  • Benefit / Cost criteria: Cost criteria are automatically inverted (10 - score)
  • Weight normalization: Raw weights are normalized to sum to 1.0
  • Score range: 0–10 per criterion per option
  • Display precision: 2 decimal places, consistently applied

Features

  • Interactive Decision Builder: Title, description, options, criteria (name, weight, type), and full score matrix
  • Ranked Results: Options ranked by weighted score with per-criterion breakdowns
  • Top Drivers: Identifies which criteria contribute most to each option's score
  • Sensitivity Analysis: Weight-swing analysis reveals how robust the winner is across ±5% to ±50% perturbations
  • JSON Export: One-click export of the full decision model and computed results
  • Shareable URLs: Base64-encoded state in URL for instant sharing
  • LocalStorage Persistence: Auto-save with 300ms debounce — no account needed
  • Preloaded Demo: "Best City to Relocate To" with 3 options and 5 criteria

Quality

  • 40 unit tests (scoring engine + input validation), all passing
  • TypeScript strict mode — zero type errors
  • ESLint — zero warnings
  • Production build — static optimization, clean output
  • GitHub Actions CI — lint → typecheck → test → build on every push/PR

Architecture

  • Local-first: All data in browser localStorage (ADR-001)
  • No backend: Zero API calls, zero accounts, zero tracking
  • Deterministic: No randomness anywhere in scoring (ADR-002)

Known Limitations

  • Single-user only (no cloud sync)
  • Limited to ~10 options × 10 criteria before UI gets crowded
  • No undo/redo
  • URL sharing has length limits for very large decisions
  • No mobile-optimized layout yet

Roadmap Excerpt

Version Focus
v0.2.0 Auth + Cloud Sync
v0.3.0 Monte Carlo Sensitivity Engine
v0.4.0 Shareable Read-Only Public Links
v0.5.0 Collaborative Real-Time Editing

Full documentation: docs/
Scoring model spec: SCORING_MODEL.md
Architecture: ARCHITECTURE.md

Assets 2
Loading