Understanding and visualizing fine-grained software modifications

点击翻译按钮获取中文摘要

Software maintenance and evolution is very expensive and knowing what changes need to be performed to complete a task takes time and experience. The objective of this research is to identify the characteristics of changes applied to software systems by their developers. This research project will define metrics to evaluate the amount of "change" performed by a modification, and create an ontology of the different types of changes that we observe.  Algorithms will be developed to automatically classify a given change within the framework of this ontology. This method of analysis will allow us to more clearly discern the relative importance of different changes in software evolution. We propose that previous studies using historical information can be improved through a weighting of the importance of certain types of changes.  Finally, we will continue our efforts to develop useful visualizations of the historical information of a software system. I believe that the research proposed herein will directly benefit software developers (and their companies) by reducing the cost (in time and money) of maintenance, and improve the overall quality of the software they produce.