Subdomain Testing

Abstract

TESTING a program is a daunting task because of the sheer size of the input domain and because of a program’s ability to store up state and use it to modify later behavior. It would be bad enough that the variety of inputs is bewildering, but when any one input may produce different results because of invisible state values, it is understandable that testing is poorly done and testers too often believe they have been thorough when they have not. There is only one way imaginable to systematize and organize test points: to group them, defining the groups as subdomains, subsets of the program input domain. Systematizing testing is important, one of those rules that engineers require to move beyond craftwork.