Highlights
- Concept models are blueprints, data models are buildings — Ross separates what the business means from how systems store it, and insists the meaning layer comes first; skipping it is why so many technically correct databases answer the wrong questions.
- Definitions are engineering artifacts — a concept earns its place through a rigorous definition anchored in the business's own language, not a developer's paraphrase; the same craft standard a concept-library holds each named concept to.
- Vocabulary + structure = the knowledge blueprint — terms alone are a glossary; terms connected by well-formed relationships (verb concepts) become a navigable structure; Ross essentially describes a typed knowledge graph before the term was fashionable.
- Data quality problems are usually meaning problems — when the same word carries three meanings across departments, no pipeline fix helps; semantic misalignment upstream masquerades as technical debt downstream.
- The blueprint is the stable layer under system churn — applications and databases get replaced, but the business's core concepts persist across generations of technology; invest the modeling effort where it compounds.
- Business rules sit on top of concept models — rules can only be expressed precisely against a vocabulary that is itself precise; Ross's stack (vocabulary → structure → rules) is the missing-third-pillar argument built from the semantics side.
Highlights on this page are generated with the help of AI.