[v9v]

The way I had understood it was that the author arrived at these problems by incrementally modifying complicated integral formulas just up to the point where Mathematica couldn't solve them, and then used the solution from Mathematica for the previous iteration, his intuition and some ad-hoc methods to make an educated guess for the answer for the "unsolvable" integral. I suppose those can be called intermediate steps, but not in the usual sense.