基于软件缺陷的可信证据度量模型

摘要

To solve the problem that the data of the trustworthy evidence measurement of software are hard to collect in the process of software trustworthy evaluation, the TEMMSD was put forward using the number and types of defects to measure the trustworthiness of software indirectly.The defects from three aspects of software developers, the third-party test and the feedbacks from users were collected and preprocessed, and the defects were analyzed and classified according to trustworthiness using the method of orthogonal defect classification and the weight of each kind of software defects was determined from the subjective and objective perspective.The software trustworthy evidence measurement was completed.An experiment based on petroleum decline tracking system was carried out.The results show that TEMMSD possesses the feasibility and effectiveness in software trustworthy evidence measurement.%为克服软件可信性评估过程中可信证据正面度量收集数据困难的问题,提出基于软件缺陷的可信证据度量模型(trustworthy evidence measurement model based on software defects,TEMMSD),利用软件系统中存在的缺陷类型、数目以及缺陷严重程度等因素,从侧面实现可信证据度量.通过软件开发者、第三方测评和用户反馈3个方面采集、获取软件生命周期中不同阶段存在的缺陷数据并对初始数据进行预处理,运用正交缺陷分类法对缺陷数据进行缺陷分析并可信归类,从主客观的角度确定缺陷类型的权重,实现软件系统的缺陷证据度量.对石油企业自然递减跟踪系统进行实例分析,分析结果表明,TEMMSD模型在软件可信证据度量中具有可行性和有效性.