The discussion centers on a proposition from O'Neill's Elementary Differential Geometry regarding the differential of a function defining a surface of revolution. The participants debate whether the differential \( dg \) can be zero on the surface \( M \) generated by revolving a curve \( C \) around an axis \( A \). They argue that while \( dg \neq 0 \) is necessary for \( M \) to be a surface, it is not sufficient, as counterexamples exist where \( dg = 0 \) on level surfaces. Additionally, the conversation highlights the importance of the function \( f \) defining the curve, noting that not all functions yielding the same level curves will maintain a non-zero differential. The conclusion emphasizes that the surface of revolution is valid, but the conditions for \( dg \) must be carefully considered.