Skip to the three cases

Early essai · Digital Construction ontology

What can an ontology say about a project?

And what must we still supply before it can explain the work required?

Start with a vocabulary, then inspect two project models. The DiCon module map, Tokyo stadium and HS2 roles keep their own contexts. Together they test a preliminary question behind Solway’s aim: an explainable work breakdown from an intended outcome, with a fair comparison to PDDL planning.

Read the sources

The Digital Construction 0.5 suite supplies terms and module imports. The project examples add selected facts, local relationships and evidence links. Their saved source hashes and retention record make those choices inspectable.

What these cases show

Imports connect definitions; they are not task dependencies. Tokyo separates objects, activities and information across several periods. HS2 separates an agent, its role and an appointment in a historical snapshot. Each arrow has a meaning to check against its source.

What is still missing

To explain necessary work we must supply the desired end state, initial and local conditions, accepted methods with preconditions and effects, and acceptance rules. These graphs do not derive work, perform new ontology inference or verify the project assertions empirically.

Continue to The work that must exist for a bounded synthetic construction that supplies those ingredients and compares ontology with executable PDDL. Its method and limits explain why the comparison establishes planning parity within that model, not ontology superiority.

Explore a model

DiCon 0.5 module map

Connections

Show node types · colour key
Ontology graphSelect a node in the searchable card list for a keyboard-accessible description of every connection.

Drag a node to pin its position. Drag the background to pan; scroll to zoom. Node size shows connection count, not influence. Solid lines are verified ontology imports; dashed lines are editorial catalogue/model relations. Use the cards for the same information without the diagram.

Node cards

Cards follow the filters. JSON contains the complete case: nodes, connections, provenance and correction notes. CSV contains all nodes and their fields, with arrays and objects encoded as JSON; it does not contain connections or case-level notes. The diagram download contains the current view.

How to read this example · corrections · provenance

Read the arrow before the picture

An ontology supplies names and constraints for concepts and relationships. A case model chooses which real or illustrative things to describe with those terms. A source supporting a fact does not automatically validate its ontology mapping.

An import makes another ontology’s definitions available. A role is something an agent bears. An appointment records a relationship between an appointing and an appointed party. The diagram layout is only an aid to reading.

To share the interactive example, use this page’s link. Discussion tools such as Loomio may remove scripts from pasted content; use a link, or an iframe only where the host permits it.

Corrections in this case

    Retention and verification source (Markdown) · Versioned ontology source hashes · DiCon 0.5 specification