These projects are for LEARNING, not production.
The goal is simple: Turn the black box transparent.
- Want to understand what React does? Build a mini version yourself.
- Want to know how frameworks work? Study SJS or Shapp internals.
- Want to learn quantum types? See them work in JavaScript first.
- Want to understand defensive programming? See real bugs from production.
We teach fundamentals. We show you what's happening under the hood.
You don't need to install a package for isEven. You don't need a framework for every project. But you SHOULD understand what those tools are doing for you when you DO use them.
Learn by building. Understand by experimenting. Master by seeing through the abstraction.
Location: sjs-framework/
A comprehensive, zero-dependency JavaScript framework designed for education. Learn modern web development patterns through clean, well-documented code.
What You'll Learn:
- Module patterns and dependency injection
- Event-driven architecture (Pub/Sub)
- Async programming (Promises, async/await)
- Cross-platform storage (IndexedDB, localStorage, Node.js)
- Client-side routing and templating
- HTTP client with caching and retries
- Performance metrics and monitoring
- Input handling and state management
Features:
- ✅ Zero dependencies - just copy and run
- ✅ Works in browser and Node.js
- ✅ 11 production-ready modules
- ✅ Complete documentation and examples
- ✅ Progressive learning path
- ✅ 7,000+ lines of educational code
Quick Start:
cd sjs-framework/examples/01-hello-world # Open index.html in browserDocumentation:
Location: defensive-compiler-design/
Learn defensive programming by studying real production bugs from a safety-critical AI compiler. Each module documents an actual bug that made it to production, the harm it could have caused, and the defensive fix that prevents the entire category.
What You'll Learn:
- Why silent defaults are dangerous (type defaulting → stack corruption)
- Pointer vs value semantics (pointer confusion → safety bypass)
- Determinism requirements (locale dependencies → audit trail corruption)
- Resource limiting (missing guards → denial of service)
- Memory optimization (unbounded allocation → OOM crashes)
- Pattern recognition across all defensive programming
- How to think like a security-conscious compiler engineer
Features:
- ✅ 5 real bugs from ARIA-026 safety audit
- ✅ Physical/emotional harm scenarios (life-critical AGI for children)
- ✅ Hands-on labs with vulnerable/safe code
- ✅ Construction wisdom applied to compiler design
- ✅ The $8 Level philosophy: Simple tools, verifiable results
- ✅ Complete manifesto on defensive programming
- ✅ Contributing guide for your own bug stories
Philosophy:
"Every bug you ship is a ladder you'll climb later."
"Silent defaults are lies."
"Boring is armor. Precision compounds."
Quick Start:
cd defensive-compiler-design/01-type-defaulting/lab make all make test # Watch the difference between vulnerable and safe!Documentation:
- 🛡️ Main README - Complete curriculum overview
- 📜 Manifesto - The defensive programming philosophy
- 🤝 Contributing - Submit your own bug stories
- 📚 Module 01: Type Defaulting - Stack corruption from silent defaults
- 📚 Module 02: Pointer Loading - Safety bypass from pointer confusion
- 📚 Module 03: Non-Determinism - Audit trail corruption
- 📚 Module 04: Recursion Limits - DoS from resource exhaustion
- 📚 Module 05: String Pooling - OOM from unbounded allocation
Who This Is For:
- Compiler/interpreter builders
- VM engineers
- Anyone building safety-critical systems
- Security-conscious developers
- Students of defensive programming
Stakes: Life-critical AGI for vulnerable children - one child harmed = game over.
Location: quantum-types-poc/
A working JavaScript proof-of-concept demonstrating Q3 and Q9 quantum number types from the ARIA programming language. Quantum types maintain TWO hypotheses simultaneously with a confidence metric, modeling uncertainty and gradual evidence accumulation—perfect for robotics, AI decision-making, and sensor fusion.
What You'll Learn:
- Superposition: Maintaining multiple hypotheses simultaneously
- Evidence accumulation: Building confidence gradually from observations
- Crystallization: Collapsing quantum state to single value when confident
- Q-functions: Conditional operations based on hypothesis states
- Gradient thinking vs binary thinking (embrace uncertainty!)
- Real-world sensor fusion (GPS + Barometer example)
- Graceful degradation when sensors fail
Features:
- ✅ Complete Q3 (3-state) and Q9 (9-state) quantum number implementation
- ✅ 8 demonstration examples showing all features
- ✅ Real-world drone altitude estimation (sensor fusion)
- ✅ Binary vs Quantum comparison with statistics
- ✅ All arithmetic operations (add, sub, mul, div, mod, sqrt)
- ✅ 7 Q-functions (qor, qand, qxor, qnor, qnand, qconf, qnconf)
- ✅ Complete API documentation
- ✅ 1,281 lines of code + documentation
Philosophy:
"Binary thinking forces premature decisions. Quantum types embrace uncertainty until evidence resolves it."
Real systems don't have instant perfect knowledge:
- Scientists don't force measurements to one value immediately
- AI doesn't commit to one path without evidence
- Humans don't make binary choices instantly
Quantum types model how real intelligence works: gradual confidence building from evidence.
Quick Start:
cd quantum-types-poc # Run 8 demonstration examples node quantum_demo.js # Run sensor fusion simulation (GPS + Barometer) node sensor_fusion_example.jsExample - Sensor Fusion:
// Two thermometers disagree: 20.5°C vs 21.3°C const temp = new QuantumNumber(9, 20.5, 21.3, 0); // Unknown which is correct // Evidence 1: Reading 21.1°C (closer to B) temp.qif(evidence_favors_b, 0, 0, 1); // Evidence 2: Reading 21.2°C (still favoring B) temp.qif(evidence_favors_b, 0, 0, 1); // Confidence built! Crystallize to single value const finalTemp = temp.crystallize(); // Returns 21.3 (sensor B won)Use Cases:
- 🤖 Robotics: Conflicting sensor readings, graceful failure handling
- 🧠 AI: Multiple hypotheses, evidence-based decisions
- 🔬 Science: Measurement uncertainty, precision tracking
- 💭 Cognitive modeling: How humans actually think (not binary!)
- ⚡ Graceful degradation: System continues when sensors fail
Documentation:
- 📖 Complete README - Full API reference
- ✅ Completion Summary - What was built
- 📁 File Manifest - Quick reference
Why This Matters: This POC demonstrates quantum types in JavaScript (familiar language) before the full ARIA language is complete. You won't find this anywhere else—it's a unique approach to modeling uncertainty that matches how real-world systems actually work!