LOD Selection

In the “Game Programming Gems” (Mark DeLoura), chapter 4.9, Yossarian King explains how to use the Magnification factor of a projection to select the appropriate LOD for an object.

“…It can be computed by transforming the object position into view space and then calculating M = xscale/zview, where xscale is the scaling parameter used in the projection equation xscreen = (xview * xscale) / zview + xcenter.”
“…hence the magnification factor M measures pixels per world unit.”

Suppose vertex C is the camera position and P is the center of the object we examine in world coordinates. How are we going to compute M? Could someone make it a bit clearer with some sample code?