Object-oriented methods create software artifacts through the application of a large number of rules. Rules are typically formulated in two-valued logic. There are a number of problems on how rules are defined and applied in current methods. First, two-valued logic can capture completely neither method developers’ intuition nor software engineers’ perception of artifact types. Second, artifacts are generally produced based only on a subset of relevant properties. Third, two-valued logic does not model explicitly contextual factors, which can affect the validity of methodological rules. Fourth, no means is supplied to deal with multiple design alternatives and to measure the quality of each alternative during the development process. High loss of information, early elimination of artifacts and process iterations are some of possible fastidious effects. To reduce these problems, this paper proposes fuzzy logic-based methodological rules. Thanks to its ability to cope with uncertainty and imprecision, and to compute with real-world linguistic expressions, fuzzy logic appears to be a natural solution for improving current methods.
Improving object-oriented methods by using fuzzy logic
MARCELLONI, FRANCESCO;
2000-01-01
Abstract
Object-oriented methods create software artifacts through the application of a large number of rules. Rules are typically formulated in two-valued logic. There are a number of problems on how rules are defined and applied in current methods. First, two-valued logic can capture completely neither method developers’ intuition nor software engineers’ perception of artifact types. Second, artifacts are generally produced based only on a subset of relevant properties. Third, two-valued logic does not model explicitly contextual factors, which can affect the validity of methodological rules. Fourth, no means is supplied to deal with multiple design alternatives and to measure the quality of each alternative during the development process. High loss of information, early elimination of artifacts and process iterations are some of possible fastidious effects. To reduce these problems, this paper proposes fuzzy logic-based methodological rules. Thanks to its ability to cope with uncertainty and imprecision, and to compute with real-world linguistic expressions, fuzzy logic appears to be a natural solution for improving current methods.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
