DevX.com has just published my article Relational Database Integration with RDF/OWL, which summarizes and updates some things I’ve written about here and in my paper for the XML 2006 conference. The article describes how to take a Eudora-based address book database and an Outlook-based one, load them into MySQL, add some OWL metadata, and then issue SPARQL queries to answer reasonably real-world questions that couldn’t have been answered without the additional metadata.
Except for the use of Outlook and Eudora to create the data models, I did the whole thing with free software that runs on both Windows and Linux. Besides MySQL, prominently featured tools included the D2RQ RDF interface to relational database managers, the SWOOP open-source ontology editor, and the Pellet OWL reasoner to perform SPARQL queries that take the OWL metadata into account when pulling answer sets.
The “databases” being integrated are pretty simple, each being a single table, but unlike some of the more obscure domains that have seen some cool ontology work, address book data is something that everyone can relate to. I hope the article can help more people see how OWL-based metadata can help apps get more out of what might otherwise seem like typical, everyday databases.