Toward a New Methodology of Software Evolution
Sabah Al‐Fedaghi · Journal of Next Generation Information Technology · 2012
Abstract Making changes to software systems often requires extensive effort since modifications propagate throughout the system, resulting in static effects such as inconsistencies, and less dynamic reliability. Studying such a phenomenon has become increasingly important and has influenced the methodology of design. Automated methods have been developed and software model evolution has been extensively studied, resulting in attempts to fix modeling semantics. This paper addresses the problem of variability at the conceptual level as illustrated in works that extend UML diagrams to allow, capture, and characterize variability in a system. The paper introduces the thesis that some efforts required when handling changes in requirements can be attributed in part to the nature of the modeling language used. Specifically, examples are analyzed to substantiate that UML diagrams lead to fragmentation, complexity, and more difficult maintenance when requirements are changed. To demonstrate this, another methodology of conceptual description is presented that provides a more focused solution.