Evaluation of Software Degradation and Forecasting Future Development Needs in Software Evolution

This article is an extended version of a previously published conference paper. In this research, JHotDraw (JHD), a well-tested and widely used open source Java-based graphics framework developed with the best software engineering practice was selected as a test suite. Six versions of this software were profiled, and data collected dynamically, from which four metrics namely (1) entropy (2) software maturity index, COCOMO effort and duration metrics were used to analyze software degradation, maturity level and use the obtained results as input to time series analysis in order to predict effort and duration period that may be needed for the development of future versions. The novel idea is that, historical evolution data is used to project, predict and forecast resource requirements for future developments. The technique presented in this paper will empower software development decision makers with a viable tool for planning and decision making.