Brief Communication Linguistic Laws and Computer Programs

Abstract

Introduction As computers become faster and cheaper, their popu-

larity continues to increase. Whatever we do and wher ever we go, we find computers. Computer programs can save time, increase efficiency, and organize our lives, but they can also prove disastrous when they fail. It is, there fore, of the greatest importance that computer programs possess such characteristics as reliability, efficiency, and maintainability. To prevent disasters, we must be able to differentiate between "good" and "bad" computer pro grams according to these characteristics. In this manner, we must be capable of measuring their quality. Unfortu nately, the application of conventional software metrics in performing this task (for example, Halstead's software science metrics, McCabe's complexity, and lines of code metrics (Come, Dunsmore, & Shen, 1986) have often not been successful (Fenton, 1991; Kokol, 1989)].

On the other hand, "infometric" research (Eghe & Rousseau, 1990) has developed a rich body of theoretical knowledge, pertinent to many real world problems. Eghe

and Rousseau (1990) has developed the concept of in formation production process (IPP) which is applicable in econometrics, demography, scientometrics, infome- trics, bibliometrics, and linguistics.

As a consequence of the weaknesses of conventional software metrics and the success of infometric research, we decided to analyze the possibility of using linguistic laws (Rousseau, 1994) in the software quality assess ment process. It is the aim of this article to present the first results and to encourage others to advance our work.

Introduction As computers become faster and cheaper, their popu-

larity continues to increase. Whatever we do and wher ever we go, we find computers. Computer programs can save time, increase efficiency, and organize our lives, but they can also prove disastrous when they fail. It is, there fore, of the greatest importance that computer programs possess such characteristics as reliability, efficiency, and maintainability. To prevent disasters, we must be able to differentiate between "good" and "bad" computer pro grams according to these characteristics. In this manner, we must be capable of measuring their quality. Unfortu nately, the application of conventional software metrics in performing this task (for example, Halstead's software science metrics, McCabe's complexity, and lines of code metrics (Come, Dunsmore, & Shen, 1986) have often not been successful (Fenton, 1991; Kokol, 1989)].

On the other hand, "infometric" research (Eghe & Rousseau, 1990) has developed a rich body of theoretical knowledge, pertinent to many real world problems. Eghe

and Rousseau (1990) has developed the concept of in formation production process (IPP) which is applicable in econometrics, demography, scientometrics, infome- trics, bibliometrics, and linguistics.

As a consequence of the weaknesses of conventional software metrics and the success of infometric research, we decided to analyze the possibility of using linguistic laws (Rousseau, 1994) in the software quality assess ment process. It is the aim of this article to present the first results and to encourage others to advance our work.