Here is a revision of my original notebook. It has much better charts at the end comparing secular energy non-conservation of the continuous, late-discretized solution versus oscillatory energy non-conservation with no secular component for the discrete, early-discretized solution.
