What Comes First, the OWL or the Bean?

The objective of this paper is to help scientific application developers who want to adopt the continuous stream of new community vocabularies  to help make their data sharable, self-describable, and easily understood. To achieve this we suggest semantic vocabulary and application integration best practices and discuss the tradeoffs of encoding vocabularies through code versus deriving code from vocabularies.