Turkish Journal of Fuzzy Systems, cilt.8, ss.73-89, 2017 (Diğer Kurumların Hakemli Dergileri)
This work is induced and inspired by the initial domain engineering activities of our ongoing project on model-driven software development automation that aims using fuzzy inference methods for a design and programming environment that is expected to provide intelligent guidance for software development. Software engineering domain contains uncertainty and imprecision that intrinsically bring about the use of fuzzy values. However the complexity and abstract nature of the concepts and decisions of the engineering domain often renders classical fuzzy logic insufficient. When the conventional fuzzy logic approach is found not to be appropriate for a certain task, switching to a more suitable t-norm/t-conorm function pair is a challenge. In some cases it may even be a necessity to extend the existing framework of fuzzy logic with other forms of knowledge representation. In this work we attempt to provide a geometrical semantics for the three common t-norm functions in order to put theory and practice into a common perspective. Probability, distance, aggregation and othersemantic metaphors are also assessed via direct practical considerations. Our work that aims faithful representation of concepts and reliable decision taking processes is hoped to contribute to the generality and scalability of fuzzy programming.