Status and Resources

help edit space_dashboard

Current Status

  • Operational: Hyperon Experimental (Rust/Python), PeTTa (Prolog), JeTTa (JVM), MeTTa-Morph (Scheme), FormalMeTTa (Scala reference)
  • Under development: MeTTa-4 semantic model, MeTTa-IL compilation pipeline, MeTTaTron F1R3FLY compiler, Windows Python packages for Hyperon Experimental
  • Proposed: PyMeTTa transpilation layer with metta-magic library, native MeTTa-to-machine-code compilation via MORK

Recent PeTTa capabilities (transcript-backed, Feb–Apr 2026):

  • Snapshot/continuation library: Crash-safe state serialization using Prolog's shift/reset mechanism and qsave_program. Enables interrupting and resuming long-running computations with atomic double-buffering (commit/commit.new pattern) to survive power outages. Directly connected to ASI Chain node execution — serialized continuations are the mechanism for checkpointing MeTTa computations on decentralized nodes. (Patrick Hammer, MeTTa Study Group, Feb 2026)
  • Vector atom space (local_vlm): Library for connecting to self-hosted LLM servers (llama.cpp). Provides chat completion API, embedding vector calculation, and nearest-neighbor retrieval from a vector-augmented AtomSpace. Demonstrates concrete neural-symbolic integration at the PeTTa level — symbolic atoms and vector embeddings coexist in the same query surface. (Patrick Hammer, MeTTa Study Group, Feb 2026)
  • External contributor growth: PeTTa pull requests are now ~80% from external contributors, up from primarily two developers. A new release is pending. (Patrick Hammer, MeTTa Study Group, Apr 2026)
  • JeTTa symbolic computation: The Kotlin-based JeTTa compiler now supports symbolic computation and has been tested on backward chaining examples. (Alexey Potapov, MeTTa Study Group, Mar 2026)

Open Problems / Research Directions

  • MeTTa-4 specification finalization — aligning multiple implementations to a common semantic model
  • MeTTa-IL optimization — proving correctness of compilation from MeTTa to various backends
  • Convergence testing across implementations (hyperon-experimental, PeTTa, JeTTa) on shared test suites
  • PyMeTTa design — defining the restricted Python subset that transpiles cleanly to MeTTa-IL
  • Module system finalization — export visibility, cross-module tokenizer imports, and a centralized module catalog (analogous to PyPI/crates.io)
  • LLM-assisted MeTTa development tooling
  • OSLF-derived spatial and behavioral types for MeTTa (Meta-MeTTa §8 future work)

Primary Sources