Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/theislampill/orthemology/llms.txt

Use this file to discover all available pages before exploring further.

Orthemology is a formal framework, not a software library. There is no package to install and no API to call. Instead, the repository ships JSON schemas, a normative registry of verdict and notation identifiers, a suite of machine-readable worked examples, and Python validator scripts that check everything against each other. This guide walks you through validating the framework’s own internal consistency, then points you to the first worked episode so you can read the framework in action.
All commands assume you are working from the repository root. The validators require Python 3, pyyaml, and jsonschema — all available via pip. No other runtime dependency is needed.

Prerequisites

Before you begin, make sure you have:
  • Python 3.8 or laterpython --version
  • pyyaml and jsonschemapip install pyyaml "jsonschema>=4.18"

Onboarding steps

1

Clone the repository

Fetch the repository and move into it:
git clone https://github.com/theislampill/orthemology.git
cd orthemology
The top-level layout you will navigate most often:
PathContents
schemas/JSON Schema files for every object type
examples/Nine machine-readable episode examples
docs/verdict-registry.yamlNormative verdict semantic IDs
docs/notation-registry.yamlSymbol table
scripts/Validator scripts
theory/orthemic-core-formalization.mdCanonical formalization
manuscript/orthemma-ortheme-systems-revised-draft.mdCurrent main manuscript
2

Validate the JSON schemas

Run the schema validator to confirm that all schemas are internally well-formed and that the normative registries agree with each other:
python scripts/validate_schemas.py
A clean run prints a summary of every schema checked and exits with code 0. If you see errors, the output identifies which schema file and which constraint failed.
The notation registry (docs/notation-registry.yaml) and verdict registry (docs/verdict-registry.yaml) are machine-enforced at this step. Symbol identifiers in schemas must agree with the registry; any mismatch is flagged here.
3

Run the full validator suite

The full suite validates every example JSON in examples/ against its declared schema, runs the deterministic fixture tests, and checks the generated project-state file against the source documents:
python scripts/validate_repo.py
The suite covers:
Seven deterministic fixtures, each witnessing a distinct cell or boundary condition in the result × pathway matrix:
FixtureWhat it witnesses
F1Nominal verification — result correct, pathway adequate
F2Correct result, defective pathway (stopped-clock case)
F3Incorrect result, adequate pathway (rare-case miss)
F4Correct result, defective pathway (defective metaorthemma — wrong reference plane, tolerance, or calibration; result happens to be right)
F5Undetermined verdict state
F6Stale-directive case
F7Safe but sub-optimal routing (V4b advisory)
Each fixture is a fully schema-valid episode record carrying an explicit verdict record whose pathway_state value is tested for correctness. These fixtures prove the framework’s four result × pathway cells are all representable — not just theoretically, but in the machine-readable format.
4

Read the pillar episode

Open examples/pillar-episode.json. This is the primary worked example — a construction-site perpendicularity check on pillar B7. It is shaped like a F1 fixture: nominal verification with a correctly bound metaorthemma, result correct, pathway adequate.The file is a JSON array of schema-instance pairs. Reading them in order walks you through the full lifecycle:
PartSchemaWhat it records
1analysis.schema.jsonThe declared analysis: task, repertoire, axes, tolerance 0.002, evidence rules, boundary
2–3metaortheme.schema.jsonTwo governing metaorthemes: the perpendicularity standard (V-component) and the calibration policy (E-component)
4orthemma.schema.jsonThe concrete case: pillar B7, version as-poured-2026-07-01
5metaorthemma.schema.jsonThe case-bound configuration token binding the standard to pillar B7, instrument level-114, and datum plane D2
6orthing-episode.schema.jsonThe episode record: actor fitter-9, time 2026-07-02T08:30–09:00Z, evidence, placement within-tolerance, route release-for-load-transfer
7claim-ledger.schema.jsonThe claim ledger: one placement claim claim-perp, its evidence references, residual burden, and closure assertion
8verdict-record.schema.jsonThe verdict record: all pathway dimensions pass, pathway_state: adequate, ROBUST_NEIGHBORHOOD: not-applicable with stated reason
The analysis declares two axes: perpendicularity (exclusive alternatives: within-tolerance / out-of-tolerance) and surface-condition (co-holding values: cracked, stained, clean). The episode’s placement records both — the pillar is within-tolerance and clean. Notice how the verdict record’s required_path field is governance-derived from the episode’s configured metaorthemes, not set by the auditor.
Once you are comfortable with the pillar episode, the other eight examples in examples/ cover different shapes: stale-directive patterns, multi-metaorthemma episodes, and defective-pathway cases.

Running the verdict semantics validator

The verdict registry enforces a strict contract between semantic IDs (machine keys such as RESULT_CORRECT, EVIDENCE_SUPPORT) and their display aliases and descriptions. To check that contract in isolation:
python scripts/validate_verdict_semantics.py
This script loads docs/verdict-registry.yaml, checks that every semantic ID referenced in example verdict records exists in the registry, and verifies that no two IDs map to the same display alias. It is also called by validate_repo.py, but running it independently is useful when authoring new episode examples.
theory/orthemic-core-formalization.md is the canonical statement of the formalization: type/token resolution, level-indexing, metaortheme and metaorthemma definitions, the verdict layer, and the abbreviation convention for analysis-relative notation. Read §1 (type/token, level-indexing) before §4 (verdicts).
manuscript/orthemma-ortheme-systems-revised-draft.md is the current full-length research draft. The abstract and §1 give the integration-discipline thesis. §2 defines analyses, orthemmata, orthemes, and profiles formally. §4 defines the verdict vector. §6 covers episode reification and multi-actor settings.
docs/decisions/ contains records 0001–0019. Decision 0001 establishes analysis-relative ground truth; 0003 establishes the result-free pathway core; 0004 freezes the verdict semantic IDs; 0005 freezes the notation registry. Reading these helps you understand why the framework is structured as it is.
The validators confirm internal consistency only — that schemas are well-formed, that examples validate against schemas, and that registries are self-consistent. They do not and cannot confirm that the framework correctly models any real-world domain. No empirical study has been run.

Build docs developers (and LLMs) love