Abstract | Several industries are using legacy softwares, developed with Structured Programming (SP) approach, that should be migrated to Object Oriented Paradigm (OOP) for ensuring better software quality parameters like modularity, manageability and extendability. Automating SP to OOP migration is pivotal as it could reduce time that take in the manual process. Given this potential benefit, the issue is yet to be addressed by researchers. This paper addresses the scenario by modeling this problem as a graph clustering problem where SP functions and function calls are vertices and edges respectively. The challenge evolving the problem is to find optimized clusters from graphs. To aid this problem, certain heuristic algorithms based on Monte Carlo and Greedy approaches are being developed. The proposed algorithms have been tested against a collection of real and synthetic data. The numerical results show that greedy algorithms are faster and produced better results than the average performance of Monte Carlo based approaches. |
---|