We defend best practices for the field of ontology engineering in support of simultaneously addressing interoperability and data quality challenges through a standardized tradecraft. Ontology engineering requires precisely making explicit the implicit semantics of data, but in practice ontology engineers frequently defer too much to domain experts, commonsense intuitions, and labeling conventions, while avoiding the modeling complexity that interoperability requires. We provide a framework through which ontologists can better understand when such deference and avoidance is useful and when it undermines the goals of the discipline. This is presented against the backdrop of historical cycles of ontology engineering, comparing past booms (such as those involving expert systems and the semantic web) with the current surge in interest due to advances in AI (such as that involving LLMs).

References

https://github.com/Applied-Ontology-Education/Ontology-Tradecraft

https://github.com/Applied-Ontology-Education/Logic-for-Ontologists-Fall-2024

https://github.com/Applied-Ontology-Education/Ontology-and-Intel-Analysis-Fall-2024