Is he right on composition? He seems to think it's about size ("small units of functionality"). Really composition vs inheritance is about dynamic versus static

Is he right on composition? He seems to think it's about size ("small units of functionality"). Really composition vs inheritance is about dynamic versus static. Acquiring the functionality at runtime or at compile time. This is explained in the book he cites. Perhaps I'm missing something.