> ## Documentation Index
> Fetch the complete documentation index at: https://www.docs.forgeshipyard.com/llms.txt
> Use this file to discover all available pages before exploring further.

# L2 · Maritime Design Ontology

> The connected model of the vessel: the answer to 'what is this ship?'

export const KnowledgeGraph = () => <Tooltip tip="Data stored as connected, typed objects (nodes and the links between them) rather than as rows in tables.">knowledge graph</Tooltip>;

export const Ontology = () => <Tooltip tip="A structured model of what kinds of things exist in a domain and how they relate. For ships: a Vessel has Hulls, a Hull has Compartments, and so on.">ontology</Tooltip>;

<Note>
  **In plain terms:** L2 represents the ship as connected engineering objects instead
  of disconnected files. It is the system's answer to "what is this vessel?"
</Note>

## What it does

L2 is the <Ontology />: a structured model of what kinds of things exist in ship
design and how they relate. It is stored as a <KnowledgeGraph />, where every
engineering object is a node and every meaningful relationship is a typed link.

The model reads naturally:

> A Vessel has Hulls and Compartments. A Compartment contains Systems. A System
> uses Equipment. Equipment cites Materials from Suppliers. Drawings reference
> Modules. Work Packages reference Modules and Materials. Evidence links to
> Drawings and Decisions.

## Why a knowledge graph instead of tables

In a traditional database you would have a vessels table and a compartments table
joined by a key. In a knowledge graph you have Vessel nodes and Compartment nodes
connected by typed links like `contains` or `adjacent_to`.

The payoff is that questions about *relationships* become simple, single-step
lookups instead of complex multi-table joins:

* "Which compartments share a bulkhead with this fuel tank?"
* "Which prior vessels have a similar system layout?"
* "Which decisions cite this rule version?"

## What L2 outputs

The canonical objects: `vessel`, `hull`, `compartment`, `system`, `equipment`,
`drawing`, `material`, `supplier`, `work_package`, `module`, `evidence`.

## How it stays trustworthy

The ontology is validated automatically; the system never accepts an object that
does not pass its rules (for example, "every Vessel must connect to at least one
Hull"). And the ontology itself, the definition of what kinds of things can exist,
is governed by **humans, never by the AI**.

Related: [Where Humans Stay in Control](/concepts/human-control).