软件产品的不断演化常常导致其质量的退化，从而增加后续维护和测试的难度。软件开发过程中实时的数据（比如，代码度量值以及bug数量等）反映了软件产品的状态变化趋势，通过这些数据可以了解软件质量的变化。然而，这些数据本身太过细节（比如，每个文件的某个代码度量值），仅仅反映这些细节数据并不能满足项目管理人员把握产品宏观变化趋势的要求。因此，提出一个基于软件开发过程数据的可视化质量监控工具，该工具不但能够基于版本库和bug库来计算每个版本的度量值以反映一个产品质量的实时变化趋势，同时，还建立缺陷预测模型，并提供数据间多维度的整合、关联、对比，来方便项目管理人员宏观地认识软件变化的各个方面。通过对一个大型开源项目的案例研究，验证了工具的有效性。结果表明，可视化产品质量监控工具能够有效地帮助项目管理人员及时发现潜在的隐患并定位问题原因。%On-going evolution of a software product usually leads to its degeneration,so that increases the difficulty of maintenance and testing tasks.The real-time datum in software development process (e.g.,the code metrics and bug numbers)usually reflects software products’change trend in their status,which could help project managers understand the changes of software quality.However,the raw data itself gets too detailed (e.g.,the code metric for each file),which might not meet project managers’need to understand the changes of software quality in a macro way.Therefore,we propose a visualised quality monitoring tool which is based on software development process datum.The tool can not only calculate the code metrics for each software (re)version based on version repository and bug repository to reflect the real-time changes trend of the quality of a software product,in addition,it also builds bug prediction model and provides multi-dimensional data integration,association and comparison between the data to facilitate the project managers having macro understanding in changes of products from different perspectives.A case study of a large open source project validates the effectiveness of the tool.Results show that this visualised product quality monitoring tools can effectively help project managers to detect potential software problems and to locate their origins.
展开▼
