Strictly speaking, there is no real need for a special set of viewing transformations. We can always compose one using a combination of rotations and translations. However, they occur so frequently that it is useful to develop a special class of transformations specifically for the purpose of mapping points from world space to eye space.