This is most useful when you are doing n-tier programming. You write a
set of classes that encapsulate your "business rules". These classes access
the database, perform whatever manipulations you need on the data, and are
able to deliver the data back to the next layer of code (typically the "user
interface"). If you wish to use a datasource to perform databinding in this
"user interface" layer, you instruct the DataSource to get its data from the
set of objects that you previously wrote for the next tier. That's why you
have this choice in "Add New Data Source".