The model as files
The whole model leaves as files, each declared in the law and written by every build from the same ground,
so what a tool reads is what the site shows.Every address below is built from ground 41e5ae2789c5; what each file is cut from and what serves it is the law's declaration of what the site projects.
The ontology and the graph, in the standards’ own forms
Load the ontology into any OWL or SKOS tool; validate the graph with any SHACL engine; run the patterns with any SPARQL engine over the triples.
- The ontology: the kinds, the relationships and the composed layer, as OWL and SKOS (Turtle)
/standards/ontology.ttl
The kinds as OWL classes and SKOS concept schemes, the relationships as object properties with domain, range and named inverse, and the composed layer as classes of its own. - Every term and every assertion, as triples with their provenance (Turtle)
/standards/graph.ttl
Every row as an individual and a concept, every assertion as a triple, every pattern, view and deliverable as an individual, and what replaced a moved address, with the build's provenance. - What a conforming graph must hold, as SHACL (Turtle)
/standards/shapes.ttl
What a conforming graph must hold, as SHACL, validated on every push by a validator that is not this build. - Each pattern as a SPARQL query over the triples
/standards/patterns.rq
Each declared pattern as a SPARQL query over the triples: one variable per role, one triple per edge.
The graph as the site reads it
The whole record as JSON, the parts an engine binds against, and the indexes a reader walks and searches by.
- The whole model, as JSON
/graph.json
The whole record: every kind, row, relationship, commitment, source, check, pattern, view and deliverable in one file, with the digest of the ground it was built from. - The parts an engine composes with, as JSON
/binding.json
The parts an engine binds against: the rows, every assertion indexed by relationship, the relationship specifications, the pattern and view declarations, the drawing rule and the operations. The nouns, the grammar and the verbs, none of the prose, about a sixth of the whole record. - The walk index: every term and the relationships between them, as JSON
/walk-index.json
The rows and the relationships a reader may walk between them, as short arrays: the steps of a walk and nothing else. - The search index: a name, an address and a kind per page, as JSON
/search-index.json
A label, an address and a kind per page: the page set without the pages. - What the site holds, for a language model (llms.txt)
/llms.txt
What the site holds, in the shape llmstxt.org asks for: a title, a summary, an account of how the site is built, then sections of links per page type, the operations, and these projections.
The schema
The shape of the graph and of every authored file, as JSON Schema, so a tool can validate what it reads.
- The shape of the JSON, as JSON Schema
/schema/graph.schema.json
The shape of graph.json as JSON Schema 2020-12, generated from the law; the schema of every authored file sits beside it under the same name.
A profile of your own
Say what your situation is on the board and it composes a profile from the vocabulary; it leaves as an application profile over the vocabulary, as a table in the standard’s columns, as one document, and as a module of the ontology, every name in it one of the addresses above.