Typically we advise that you don't integrate your domain model through to a relational DBMS in the early exploration/prototyping stages of development, because it's then easier to refactor and rework the model as your understanding of the complexities of the domain deepens. During this time, just use the in-memory object store and use fixtures to populate the object store for demoing. (You might also want to check out the FitNesse integration provided by the Tested Objects sister project).

However, there will come a point (eg after a couple of iterations) when the domain model is stabilizing. At this point you'll want to start integrating your new/updated domain objects into an RDBMS. The JPA Objects' Maven archetype is designed to help you complete this integration work.

When run the archetype creates a parent module along with some child modules, the idea being that you incorporate these modules into your own parent module (by editing the <modules> element). However, because most of the complexity involved is in adding the JPA annotations, what the archetype also does is supply examples of annotated domain objects for the "claims" applications that ships with Naked Objects. What you might therefore want to do is to gain experience by running the Maven archetype against this claims application. When you've seen how it works, you can run it again for your own application.

The following sections provide a workthrough of how to run the Maven archetype and show how to get the claims application running with it.