ABSTRACT
At this point in the book we’ve subjected the conventional software system development methodologies to scrutiny and exposed their weaknesses, especially with respect to the possibility of engineering Al software. We’ve also taken a brief look at a few proposals for schemes, or new paradigms, for overcoming certain of the weaknesses that are apparent; they do not solve, or circumvent, the fundamental problems, I hope that I’ve convinced you of that as well. And finally, I’ve been promoting the need for an iterative, evolutionary approach to engineering Al systems. The RUDE cycle was presented, but within it we could see even worse problems for the possibility of engineering robust, reliable and maintainable software systems. But in the absence of serious schemes that do solve the important problems, I propose to persist with the RUDE cycle and to examine its potential as a basis for Al-software development.
