Semantic Modeling for Data

Chapter 2. Semantic Modeling Elements

Take advantage of the ambiguity in the world. Look at something and think what else it might be.

Roger von Oech

A difficulty you might face when talking with other data professionals about semantic models is understanding and agreeing on what these models might look like and what elements they might consist of. If you ask a database developer who has been working all their life with relational database systems, their answer will most likely include tables, fields, and primary and foreign keys. If, on the other hand, you ask an ontology engineer with a background in Semantic Web technologies, they will mention classes, object properties, datatype properties, and individuals. And if your interlocutor’s origins are traced back in linguistics, it’s quite probable that they will refer to synsets, lemmas, synonyms, and hyponyms.

In practice, all these people refer to similar types of semantic modeling elements, yet they are accustomed to different terminologies that make it hard to immediately comprehend one another. At the same time, different people and communities use the same terms for different elements, making common understanding even more difficult.

Such a common understanding, though, is crucial for building data models that are semantically interoperable to each other. Until we all have at our hands a universal semantic modeling language that everyone uses (which I don’t think will ever happen), we need to be able to draw conceptual ...