Comments on P.D. Drummond's central partial difference propagation algorithms