This paper improves a program development method proposed in a recent paper by R. G. Dromey. It is argued that writing the postcondition in a judicious way can lead to programs which are easy to verify and efficient. Furthermore, we want to investigate feasibility of bridging the gap between the art of programming and the science of programming. It might be that the art is in writing the postcondition and the science in developing an algorithm from the postcondition. Examples are given to illustrate Dromey's method and improvements to it.
Proceedings of the ACM 30th Annual Southeast Regional Conference (ACM-SE 30), Raleigh, North Carolina, United States, 08-10 April 1992,