The model of derived services presented does not map consistently for all types of data resources
E.g. you may create an empty file and the add data to it but you create a new database table and interact with the 
database to insert data
This feels ok but we need a clear way of documenting expected use patterns. We need a new primer. 
.