My very good friend Elias Torres is (and has been) thinking about RDF. That is, he's been thinking about the seamingly endless back and forth debate between members of the RDF camp and members of the XML Syntax camp over which approach to producing interoperable and extensible data formats is best. For some, the Data Model is king. For others, Syntax is king. The Data Model folks are willing to sacrifice a bit in syntax efficiency because it's the data model that is most important. The XML Syntax folks tend to take a the-syntax-is-the-data-model approach works towards creating as efficient a syntax as possible. What is interesting is that both camps are generally made up of the people who are producing the software and not the people who are consuming it. People who are consuming the software believe that the data is more important than the data model or the syntax. The paradox is that those consumers want a data model that properly represents the data they're looking to access and a syntax that allows them to get to that data efficiently and without jumping through a lot of hoops but they never actually ever want to have to think about the data model or the syntax. What consumers want is to build applications to solve specific problems. They just want the underlying technology to allow them to do that without getting in the way.