application_development

To get full benefits from OWL (here I refer to the OWL2+SWRL bundle), we need a full integration of OWL inside applications. My approach goes even further: I consider OWL as a programming language for developing applications. The application schema will be replaced with a modular ontology, and the business rules will be implemented as OWL axioms. An application manager (in a procedural programming language, for example, Java, C++, depending on OWL API availability) will run OWL modules in a reasoner, and perform CRUD operations on them.

The benefits are enormous. The application manager is very simple, so there is no much of feeding grass for bugs. Of course, inserting bugs in the OWL code is possible, but those bugs are easy to detect. It is easy to change the schema dynamically, which would be a disaster for relational databases.

This approach is not performance motivated. Depending on application requirements, performance can be better or worse. There are numerous ways for performance improvements in the OWL technology.