Knowledge representation
The umbrella concept of this spoke: describing meaning in a form machines can store, exchange, and reason over. In practice that means the semantic web family — a layered stack built on a single graph data model.
The stack
- Data model — rdf: everything is a triple (subject–predicate–object); triples form a graph. An abstract model with interchangeable serializations (Turtle, JSON-LD, RDF/XML, N-Triples).
- Schema / ontology — RDFS then OWL add classes, property hierarchies, and reasoning; SHACL adds validation. (Named here; not yet separately paged.)
- Query — SPARQL, the graph query language for RDF.
- Vocabularies & deployment — shared term sets like schema-org, and knowledge bases like Wikidata and DBpedia, are where the model meets the real web.
- Authoring — building ontologies needs tooling (Protégé-class editors); see ontology-dev-tools.
The defining tension
The heavy, reasoning-oriented end of the stack (rich OWL ontologies, automated inference) has chronically under-delivered on adoption, while the shallow, pragmatic end — schema-org structured data for search — won at internet scale. The recurring lesson (see synthesis): adoption follows a concrete payoff and a low authoring cost, not expressive power.
Reflexive note
schema-org is the @type vocabulary this hub itself runs on — every spoke’s pages are
schema.org-typed JSON-LD. This spoke documents the model the wiki is built on.
Boundary
Distinct from the sibling research-wiki (markdown PKM / tools-for-thought + formal-methods theorem proving, and its knowledge-graph node) — a different tradition; cross-link, don’t duplicate. RDF knowledge graphs feeding LLMs (RAG grounding) bridge to agentic-tooling-/llm-providers-wiki.
Related
rdf · schema-org · ontology-dev-tools · knowledge-graph · synthesis