Dion is writing about where to put config files.

I always try to make it possible to do as much work as possible without leaving the IDE. For example, I have something I call "standalone mode" where I create a simple StandaloneServer class with a simple main method which uses the Jetty API to start up the server directly inside the IDE. For this reason I always keep my config files around in such a way so that "standalone mode" works without requiring the use of a build script. That is I "Mimic The Deployment Structure". In this mode I try to stub out high-end J2EE services (not available in plain vanilla Jetty) such as JTA, JNDI, JAAS and so forth. Using Spring Framework this is easily done. I also try to stub out backend services that we don't have reliable connectivity to.

The released app is usually deployed to a real appserver where JTA would take care of committing transactions and JNDI would be used to look up pooled DataSources. Therefor I have a build script that converts the standalone mode deployment structure into a deployed production mode ear file, tweaking config files and so forth to make it behave well in a real appserver.

Doing it this way you can have the best of both worlds. You get high productivity in development because you don't need to do a full compile-build-redeploy-cycle every time you've done a change (restarting Jetty takes about 3 seconds), in fact you should never need to leave the IDE to execute build scripts during normal development. You also get an app that is a good J2EE citizen, demarcating transactions with JTA, using JNDI for DataSources and so forth. It really does wonders for productivity!