Research annex series
Research Annex · Keet Notes
Study notes from Maria Keet's ontology-engineering textbook, applied to the operating layer Hadto builds for small businesses.
This series is source work, not the reading path. If you run a business, start here — or browse the rest of the research annex.
- Part 1 Keet notes: the reading guide
Twelve study notes from Maria Keet's ontology-engineering textbook, in book order, with the three consolidated chapter notes marked. Start here if you are entering the annex.
- Part 2 Start with the business problem, not the schema
A tidy schema is not the goal. The model has to help an operator make decisions, hand off work, and run the business.
- Part 3 Reasoning is a control surface, not an academic luxury
Reasoning is not garnish. It is part of how a business system proves the model still makes sense.
- Part 4 Description logics tell us what to automate and what to bound
Formal models matter when they tell the team what the system can infer, explain, and afford to compute.
- Part 5 The ontology stack needs explicit contracts
OWL is only one layer in a broader stack. If a company does not name the boundaries between data, validation, and reasoning, operators are left guessing.
- Part 6 Keet notes, chapter 5: methodology and the quality gates
Five study passes through Keet's methodology chapter converge on one lesson: consistency checks stop at the class layer, and the expensive failures live everywhere else.
- Part 7 A venture platform needs a stated ontology posture
A shared upper layer only creates leverage when the company states what it means, what it governs, and where it intentionally stops.
- Part 8 Not every `contains` edge is parthood
Business systems get dangerous when every contains-style relation is treated like structural part-of.
- Part 9 Keet notes, chapter 7: bottom-up sources need lifting and review
A database schema is not the business model, an AI label is not accepted knowledge, and a spreadsheet column is not a concept. Chapter 7 is about the review that stands between source and ontology.
- Part 10 Your ontology platform has to say where the live data lives
Operators should not have to guess whether the ontology stores the truth, points to it, or only documents intended structure.
- Part 11 Multilingual ontology work is governance, not extra labels
If Hadto wants ontology infrastructure that can travel across communities, multilingual support has to be treated as a governed operating layer—not an afterthought of translated strings.
- Part 12 Confidence scores are not fuzzy semantics
Scores can help with review, but they do not change crisp semantics unless the platform says exactly what they mean.
- Part 13 Keet notes, chapter 11: when is ontology work published?
Green checks are a layer, not a verdict. Chapter 11 is about the release discipline: readable trust, an inspectable research intake, and derivative surfaces that admit they are derivative.