Rule-Based Programming is a broad presentation of the rule-based programming method with many example programs showing the strengths of the rule-based approach. The rule-based approach has been used extensively in the development of artificial intelligence systems, such as expert systems and machine learning. This rule-based programming technique has been applied in such diverse fields as medical diagnostic systems, insurance and banking systems, as well as automated design and configuration systems. Rule-based programming is also helpful in bridging the semantic gap between an application and a program, allowing domain specialists to understand programs and participate more closely in their development. Over sixty programs are presented and all programs are available from an ftp site. Many of these programs are presented in several versions allowing the reader to see how realistic programs are elaborated from back of envelope' models. Metaprogramming is also presented as a technique for bridging the semantic gap'. Rule-Based Programming will be of interest to programmers, systems analysts and other developers of expert systems as well as to researchers and practitioners in artificial intelligence, computer science professionals and educators.

Rule-Based Programming is a broad presentation of the rule-based programming method with many example programs showing the strengths of the rule-based approach. The rule-based approach has been used extensively in the development of artificial intelligence systems, such as expert systems and machine learning. This rule-based programming technique has been applied in such diverse fields as medical diagnostic systems, insurance and banking systems, as well as automated design and configuration systems. Rule-based programming is also helpful in bridging the semantic gap between an application and a program, allowing domain specialists to understand programs and participate more closely in their development. Over sixty programs are presented and all programs are available from an ftp site. Many of these programs are presented in several versions allowing the reader to see how realistic programs are elaborated from `back of envelope' models. Metaprogramming is also presented as a technique for bridging the `semantic gap'. Rule-Based Programming will be of interest to programmers, systems analysts and other developers of expert systems as well as to researchers and practitioners in artificial intelligence, computer science professionals and educators.

A small program is presented to motivate the concerns for programmer productivity and program quality that are the central issues of this set of essays. The example is one which demonstrates the performance aspect of programming. In order to achieve program quality, where a program is understood and known to be correct, we need a primary program description. This primary program description not only describes the program but is also used to generate the program. The method of applying primary program descriptions to produce programs is called metaprogramming and is described in Chapter 3. In the later chapters, we show how the method can be analyzed from an economic point of view to address the issues of productivity as well. 1 Introduction In thinking about programming over the last decade, I have concluded that very little is known about the process of programming or the engineering of software [1]. The consequence of having very little established truth to use as a basis for thinking about programming is that almost every conclusion must be reasoned out from first principles. Also, you cannot rely solely on textbooks but must use experimentation and direct observation to gain some experience with which to proceed.