Martin F., the computer scientist: bald with a short grey beard and glasses, a mustard T-shirt under a navy hoodie, a handheld engineering tool in one hand.
Martin F.
the computer scientist

Ontoledgy · AI context engineering

The AI Ontology
Context Library

Connect the data we hold to the world it describes. Make meaning explicit, then test it.

Data Meaning World
Richard F., the 4D ontologist: tall and lanky, wild grey hair, open white lab coat, spacetime goggles pushed up into his hair, a slim handheld device in one hand.
Richard F.
the 4D ontologist
New here? Meaning connects data to the world. An ontology models what exists and how things relate. Test it as context for AI.

01

The library, mapped

Everything here sits on one piece of terrain: a data side, a world side, and the bridge of meaning between them. The journey walks that terrain end to end — and turns it on the machine that mimics us; the skills staff it. There are three ways in — start as a data architect with the models, a finance modeller with the forecast, or an accountant with the books — and all doors open on the same bridge. Follow the spine, or jump straight to any part.

Three ways in · pick a door

Part 2 of 5 · guided World side Cross over →

From Words to the World

A definition leaves the leap from word to world to a brain; an ontology carries the world's structure across with it. An exposé on what that structure is — the six hard questions every ontology must settle, and the indexical plank that seats an agent in it.

Part 3 of 5 · guided World side See the machine →

The Semantic Pipeline

You don't author the ontology by hand — you reverse-engineer it from the data models you already own and evolve it as each new source arrives. The five-stage bCLEARer pipeline, with the upper ontology as its machinery.

Part 4 of 5 · guided · interactive The grounding Ground the generator →

Grounding the Generator

The LLM mimics the brain that did the word→world leap — but it lacks a deep long-term memory and Aristotelian common sense. The grounded ontology supplies both; an interactive matrix maps which AI pattern draws on which grounding mechanism, with deterministic identity and 4D time as the differentiators.

Part 5 of 5 · guided · interactive The architecture See it whole →

Putting It All Together

The whole reference architecture on one canvas — many source systems in, each a thin slice through C·L·E·A·R; every slice's Assimilate hydrates one shared ontology; every Reuse projects it two ways — founded datasets for the target systems, and the graphs that ground the LLM.

Companion · interactive Staffs the terrain Explore →

The Skills that do the Work

The skill library mapped onto the same terrain — ontologists work the world, engineers the data, architects the bridge, orchestrators drive the lifecycle. Expandable diagrams for the facet taxonomy, inheritance trees, scope × language grid and the orchestration stack.

Companion · reference Under the hood Open →

The Technical Components

The reusable mechanisms the pipeline is built from, each opened up on its own page — starting with the deterministic identity scheme that merge and grounding turn on. b-units, Universes and orchestration follow.

Companion · reference Every term Look up →

The Glossary

Every term the journey defines, gathered and cross-linked in one place — from data models and ledgers, across the bridge to ontology, and through the bCLEARer pipeline. Search it, filter by where a term comes from, or jump by letter.


02

Notes & plans

Working documents on GitHub — the plans behind the skill families and the modelling glossary.