Computing program modularizations using the k-cut method

The problem of producing modularizations of huge legacy software systems which lack an obvious modular structure is an important one. Such modularizations allow maintenance and reuse of smaller, more manageable pieces of program source code, and also may provide insights into the overall structure of the software system. We consider the application of an algorithm from graph theory known as the k-cut method to the domain of computing program modularizations. We describe the k-cut method, its associated algorithms, and the problem's application to reverse engineering of legacy systems. We also describe our experience in applying the method.