I am starting to construct more abstract classes and wonder how people are deciding to choose the source of parent for their new classes. If I use a subclass of custom in a 3 tier class hierarchy I would have two choices. If I create the new class in the third app tier, I could subclass either the custom class in the third tier or I could subclass the parent of the custom class in the third tier. I realize there is not much difference here, but since this activity seems that it could become addictive it would be nice to have a consistent method that had some reason to it. Subclassing the third tier custom class actually gives one level more of specialization potential. Is that why it is the best?
TIA