odl-kernel 0.9.1

ODL (Organizational Definition Language) Execution Kernel & Physics

ODL Kernel

The Physics Engine for AI Agents.
Deterministic Foundation for Probabilistic Systems.

---

The Deterministic Execution Engine for Organizational Definition Language (ODL).

odl-kernel is the reference implementation of the ODL "Physics." It is a stateless, pure-functional execution engine designed to manage the lifecycle of autonomous AI agents and human-AI collaboration.

Unlike traditional workflow engines, the Kernel acts as a rigorous state machine that derives the next organizational state and necessary side effects (Commands) based on a provided JobSnapshot and KernelEvent.

⚛️ Core Principles

• Functional Core: The Kernel is side-effect-free. It does not access databases or external APIs directly; it only calculates the transition.
• Deterministic Identity: Every node is assigned a mathematically derived UUID v5 based on its structural path. This ensures that process IDs remain consistent across restarts, replays, and forks.
• Deep Analyze: The engine performs fixed-point iteration to advance states until a stable equilibrium is reached, allowing the host to observe only meaningful state changes.

Usage

The integration follows a strict "Load -> Analyze -> Apply" pattern.

import odl_kernel
from odl_kernel.types import JobSnapshot, KernelEvent

# 1. Prepare Input (Stateless)
snapshot = JobSnapshot(job=job_data, nodes=node_map)
event = KernelEvent(type="TICK", occurred_at=datetime.now())

# 2. Analyze (Pure Physics)
# The Kernel calculates the next state and required commands.
result = odl_kernel.analyze(snapshot, event)

# 3. Apply Output (Side Effects)
# The Host runtime persists updated_nodes and executes result.commands.
my_db.save(result.updated_nodes)

📖 Deep Dive: The Physics of Execution

The code above is just the entry point. The Kernel is a stateless, pure-functional "Physics Engine" designed to run inside your application.

To understand the architecture (Functional Core / Imperative Shell) and how to implement the Host Runtime (Database persistence, LLM API connections, and the Event Loop), please read the full implementation guide:

USAGE.md: The Host Implementation Guide

🏗 Relationship with odl-lang

This repository provides the execution environment. To define the organizational structures and compile them into an executable format, use the odl-lang library.

🌹 Dedication

Dedicated to the memory of Prof. Kazuhisa Seta, and to all researchers who continue to push the boundaries of Knowledge Engineering in his lineage.

⚖️ License & Commercial Terms

Copyright (c) 2026 Centillion System, Inc. All rights reserved.

• Licensor: Centillion System, Inc. (https://centsys.jp/)
• Contact: [e-mail]

This software is licensed under the Business Source License 1.1 (BSL 1.1).

Additional Use Grant

You are authorized to use this software in a production environment under the following conditions:

1. Managed Service Restriction: You may not use this software to provide a competing "Managed ODL Service" to third parties.
2. Revenue Limit: Use is free for individuals, academic institutions, and organizations with annual gross revenue less than $10,000,000 USD.
3. Transition to OSS: This software will convert to Apache License 2.0 exactly two years after the release date of each version.

Organizations exceeding the revenue limit or wishing to provide managed services must obtain a commercial license from Centillion System, Inc.

---

The deterministic ID generation logic in this Kernel uses a proprietary signature for provenance tracking and identity verification.