Skip to content

REGISTRY.adoc

Latest commit

 

History

History
100 lines (81 loc) · 4.76 KB

REGISTRY.adoc

File metadata and controls

100 lines (81 loc) · 4.76 KB
Raw

Database and Language Registry

Table of Contents

This is the authoritative ownership map for the portfolio. It records semantic pairings, not a claim that every language-to-database path is implemented or proved. nextgen-databases coordinates; implementation belongs to the owners below.

Families

Family Canonical repositories Language purpose

VeriSimDB / VCL-UT

verisimdb, vcl-ut

Consonance and admissible transitions, including evidence and usage conditions.

Lithoglyph / Glyphbase / GNPL

lithoglyph, glyphbase, gnpl

Narration/projection: accounts, stances, warrants and rival interpretations.

QuandleDB / KRL

quandledb, krl

Knot construction, transformation, resolution and retrieval of relevant evidence.

Vocarium / Hermeneia

vocarium, hermeneia

Situated meaning and sufficiency under a declared use model.

GNPL’s existing typed selection/storage machinery is private implementation code. Historical source namespaces are compatibility identifiers, not separate public languages or a fixed narration-lowering contract. VCL-UT and KRL likewise retain their semantic identity even where an implemented fragment uses familiar access syntax. No additional public language is precommitted for generic reporting.

Independent knot projects

Project Role and relationship

Skein.jl

Independent Julia knot storage/indexing and invariant library, used by QuandleDB.

KnotTheory.jl

Independent Julia mathematical toolkit for diagrams, braids, invariants and transformations, used by QuandleDB.

Tangle

Separate Turing-complete language for knot mathematics. Shared subject matter; no KRL compilation dependency or inherited proof guarantee.

Supporting research

NQC remains separately tracked calculus research. Its historical directory is not evidence of an additional language required by each family. Extraction status remains governed by the migration documents below.

VeriSim satellites

Component Disposition

verisim-core/ (slim Julia identity-core)

→ own repo hyperpolymath/VerisimCore.jl (Julia-package pattern); create + push

verisim-modular-experiment/

→ folded into hyperpolymath/verisimdb as a standard feature (no longer an experiment)

Reference strategy (after extraction)

Every repo is referenced by link from this registry. There are no git submodules.

This previously read "`verisimdb` and lithoglyph are git submodules of this coordination repo". That was never true of either: git ls-files --stage showed both as ordinary tracked files, not gitlinks, and no .gitmodules has ever existed here. Corrected in favour of what the repo does rather than adding submodule machinery to make a stale sentence true — a submodule would have to be fetched with submodules: recursive in every workflow that touches the tree, which is real cost for a coordination repo that only needs to point at its members.

lithoglyph/ and verisimdb/ now hold pointer READMEs only; their 819 and 713 files were removed once each extraction completed (2026-07-27 and 2026-08-03).

What belongs in this (coordination) repo

  • Portfolio docs: README.adoc, TOPOLOGY.adoc, ROADMAP.adoc, EXPLAINME.adoc

  • This registry (REGISTRY.adoc)

  • Cross-database integration tests (tests/)

  • Shared infra & environment (flake.nix, guix.scm, Justfile, stapeln.toml, opsm.toml)

  • Governance & metadata (.github/, .machine_readable/, .well-known/, LICENSES/, CONTRIBUTING.adoc, SECURITY.adoc, 0-AI-MANIFEST.a2ml)

Note

The per-database directories still in this repo are legacy and are being extracted — see docs/migration/RESITE-DATABASES-TO-OWN-REPOS.adoc. Do not add new content to them.

lithoglyph/ and verisimdb/ are already done and now fail the placement guard rather than warning. quandledb/, nqc/, typeql-experimental/, verisim-core/ and verisim-modular-experiment/ remain grandfathered.

verisimdb was a fork, not a duplicate. Do not assume the next extraction is a clean subset of its own repo — verify byte-identity per file first, and port whatever is unique before deleting.