Learning Resources

help edit space_dashboard
Draft — This content has not been approved for publication.

← Back to MeTTa Programming Language

Status: Current. The primary structured learning resource is the mettatraining curriculum developed by iCog Labs, covering MeTTa from foundations through cognitive architectures in 51 lessons across 6 modules. Additional learning materials include examples and test suites distributed across implementation repositories.

mettatraining Curriculum (51 Lessons)

The mettatraining portal (MkDocs site with MeTTa challenges and solutions) provides a structured path organized in 6 modules:

Module 1: Foundations of MeTTa (Lessons 1–6)

Environment setup, basic syntax, built-in data types, variables, function definitions, control flow, loops, AtomSpace fundamentals, pattern matching, and the standard library. Begins with functional programming prerequisites (recursion, immutability, higher-order functions).

Module 2: Advanced Logic and Data Structures (Lessons 7–11)

Module management and file manipulation, recursive logic, higher-order functions, data structures (types, constructors, lists, trees, sets, maps), state monad, and Python bindings with mutable structures.

Module 3: Building and Non-Determinism (Lessons 12–18)

Practical revision exercises, guided builds, non-deterministic logic (introduction and applied), AtomSpace space operations review, and custom/nested atomspaces.

Module 4: Integration and MORK (Lessons 19–27)

Common AtomSpace design patterns, unification and self-rewriting code, Python integration (custom function wrapping, advanced patterns), debugging techniques, and introduction to MORK (architecture differences, environment setup, pattern matching queries, MM2 architecture, code stepping).

Module 5: MM2 Deep Dive (Lessons 28–39)

MM2 fundamentals (sources, sinks, priority), set operations, control logic (basic through advanced), macros (definitions, implementation, advanced), and large-scale program architecture and optimization.

Module 6: Cognitive Architectures and Conclusion (Lessons 40–51)

Large-scale program scaling and optimization, introduction to Hyperon (symbolic vs. neurosymbolic AI, cognitive architectures), and a survey of core algorithms: OpenPsi, ECAN, Pattern Miner, MOSES, PLN, and cognitive synergy. Concludes with ecosystem overview (alternative compilers including PeTTa, ASI Chain and MeTTa-IL) and development mentoring.

Additional Resources

  • metta-lang.dev — Official MeTTa language documentation site
  • hyperon-experimental — Reference implementation with extensive docs directory: formal grammar (MeTTa language specification), minimal MeTTa specification (assembly-level semantics), module system developer guide, and stdlib reference
  • metta-examples — Comprehensive MeTTa language feature demonstrations
  • metta-testsuite — Multi-interpreter test suite for cross-implementation validation
  • MeTTa-AI-Assistant — RAG-based MeTTa coding assistant (FastAPI + React)
  • MeTTa challenges: 14 programming challenges with 6 published solutions in the mettatraining repository

Related cards: MeTTa Full · Hyperon Experimental · MORK · PLN



Discussion