Genericity in Geometric Computing

By choosing a type such as double to represent the coordinates of a point, the Point class and all accompanying classes are forever sealed at that precision. Yet another application may need to use an identical set of classes at a different precision, perhaps float, long double, or a rational number type from the GMP library such as mpq class. This chapter discusses how the genericity of C++ can be used to design classes that delay the decision of the number type used for coordinates. Such a delay increases the chance that multiple systems could be developed based on the same set of foundational geometric classes.

It is best to read Chapters 6 and 7 simultaneously, perhaps by reading either of the two chapters before and after the other.