Programming to Interfaces

The basic technique for implementing dependency inversion using interfaces is called PROGRAMMING TO INTERFACES, and it’s always a good way to protect yourself (and your classes) from change. I’ve actually seen it described as “a fundamental principle of OOP”. It’s a good practice, but I personally doubt it’s fundamental, not in the way that, say encapsulation is.

The thing to be careful of with programming to interfaces is a cluttered class hierarchy. You should beware of any interface that is only implemented by one or two concrete classes. It’s probably just added complexity (otherwise known as “noise”).

This Is How Things are Usually Done

The Safe Way