there's probably a rational category based on the size/complexity of the data model. I get the sense that very complicated data models are what really ruins the agile approach for complex environments.