A typed approach to ontology manipulation (experience report)
summary
Presents a practical pipeline that uses the statically typed XML language CDuce to parse RDF/XML, convert the Europeana fashion thesaurus into OWL-style classes, and merge that with a people/jobs ontology to model 12th‑century dress and social roles. Types (intersection, sequence) enforce structure at compile time, pattern‑matching generates new tags like man/woman and a madeOf relation, and Protege is used for manual massaging where CDuce gets verbose; the result is a working, richer merged ontology.
AI-written summary · about
chapters (15)
- 0:00Introduction
- 0:41What is an ontology
- 2:53T-box and A-box in OWL
- 5:51Focus on XML for ontologies
- 7:39Introducing Seduce
- 10:36Seduce syntax and types
- 13:30Functions and overloaded types
- 16:43Intersection types and matching
- 18:29Transforming XML with Seduce
- 21:42Ontology type in Seduce
- 23:53Porting a thesaurus
- 26:24Automating transformations
- 28:12Merging ontologies
- 30:19Integrating ontologies
- 31:01Conclusions and future
related talks
31mBuilding a Scalable, Reproducible Ontology System to Improve Deep Learning Data
39mSharing Data, hiding Complexity (with Clojure and RDF)
34mProfile driven transformation service at the Dutch Central Bank
29mBuilding a knowledge graph with the Natural History Museum to work as a taxonomic equivalence engine
11mIntroducing an Enhanced Metadata Broker for Manufacturing Data Spaces
39mAdventures in Understanding Documents
14mOntology-Backed GraphRAG: Injecting Biomedical Logic in LLMs for Drug Discovery
41mDatomic Auto Schemas