Abstract
The authors argue that strong mutation testing and weak mutation testing are in fact extreme ends of a spectrum of mutation approaches. The term firm mutation is introduced to represent the middle ground in this spectrum. The authors also argue, by means of a number of small examples, that there is a potential problem concerning the criterion for deciding whether a mutant is dead or live. A variety of solutions are suggested. Practical considerations for a firm-mutation testing system, with greater user control over the nature of result comparison, are discussed. Such a system is currently under development as part of an interpretive development environment.<>