The primary task in solving a physics problem is to select ways of viewing the problem in terms of physical systems whose behavior is described by physical laws. The physical systems are, in general, only approximate models of the real-world systems. As models of real-world systems are made more accurate, the equations involved quickly become unmanageable; furthermore, certain special cases of physics problems which frequently occur can be solved using greatly simplified equations. Success in solving physics problems therefore depends crucially on selection of physical systems which satisfactorily model the real-world systems and which possess tractable mathematical models. Selection of the physical systems used as models is generally based on qualitative features (including ranges of numerical magnitudes) of the real-world systems. Many problems are best solved by considering the same real-world system from multiple viewpoints and relating the viewpoints to each other, often by identification of their components. (Author)