Files
LithosAnanake/docs/working/archive/quality

Quality Assurance

This directory contains testing, validation, audits, and quality assurance processes for StarForth.

Quality Processes

  • audits/ - Code audits, implementation reviews, codebase audits
  • phase-tracking/ - Phase completion checklists and session summaries
  • regression/ - Regression detection framework and reports
  • validation/ - Validation protocols and comprehensive physics validation

When to Use

Before Committing:

  • Run make test - All 936+ tests must pass
  • Run make with -Wall -Werror - Zero warnings required

Before Releasing:

  • Run full validation suite
  • Execute DoE mode: ./starforth --doe
  • Verify 0% algorithmic variance

After Major Changes:

  • Run regression detection framework
  • Compare DoE results with baseline
  • Update validation documentation

Phase Completion:

  • Create implementation audit
  • Run comprehensive physics validation
  • Document results in phase-tracking/

Test Coverage

Unit Tests (POST - Power-On Self Test)

  • Q48.16 fixed-point arithmetic
  • Inference statistics (ANOVA early-exit)
  • Decay slope inference (exponential regression)

Dictionary Tests

  • FORTH-79 word compliance across 18 modules
  • Stack words (DUP, DROP, SWAP, ROT, etc.)
  • Arithmetic words (+, -, *, /, MOD, etc.)
  • Control words (IF, ELSE, THEN, DO, LOOP, etc.)
  • Defining words (:, ;, CREATE, DOES>, VARIABLE, CONSTANT)

Integration Tests

  • VM + physics subsystems
  • Heartbeat coordination
  • Block I/O subsystem

DoE Validation

  • Gold Standard: 0% algorithmic variance across 90 experimental runs
  • Deterministic behavior guarantee
  • Reproducible across platforms (Linux, L4Re)

Quality Metrics

Current Status:

  • 936+ tests passing
  • 0% algorithmic variance (formally validated)
  • Zero warnings with -Wall -Werror
  • Strict ANSI C99 compliance

See Also

  • Comprehensive validation protocol: validation/comprehensive-physics-validation.md
  • Break-me testing report: validation/break-me-report.md
  • Regression detection framework: regression/detection-framework.md