Skip to main content

​
Notation Sets, Living Lexicon, and Cross-Domain Semantic Index

The notation-sets proposal establishes a canonical, continuously evolving notation vocabulary supporting Grimoire compression (#309 / #182) and cross-domain semantic indexing. It separates canonical notation from domain aliases and domain-specific syntax, connects the vocabulary to the repository’s existing pointer/concept/tool/context indexes, and defines a research-driven evolution loop so the glossary stays synchronized with implementation and operator governance.

​
Status

  • Proposal status: posted (not accepted) - execution remains gated by repo-gate + termux-smoke.
  • Malleable by design: revisions are additive and attributed via the evolution ledger (NSE-004 loop, NSE-008 ledger). This is a revision pass, not a freeze.
  • Latest revision: PR 390 - formalized category-theoretic notation, cross-domain mappings, and proposal/test alignment while continuing the living-lexicon research line.

​
Research basis

This revision integrates two prior research sessions:
  • Language information density & ADLM - Coupé et al. 2019 (cross-language convergence near ~39 bits/s); Petrov et al. EMNLP 2023 (tokenizer inequity across scripts); FLORES-200; tiktoken; Adaptive Dynamic Language Mixing as a tokenizer- and model-specific semantic codec with a canonical-IR truth layer.
  • Grimoire compression & category-theoretic IR - layered compression stack (zstd, LZ4, Brotli, FastCDC, xdelta3, CBOR, Parquet); canonical IR for category theory; per-corpus dictionary training.
See NSE-009 to NSE-021 in the full proposal.

​
Item index

#320, #309, #182, #175, #126, #304, #196, #177, #208, #274

​
Source

Full proposal: docs/proposals/active/notation-sets/ (ITEMS.md, MANIFEST.md) and registry entry in docs/proposals/registry.yaml.