软件可信评估研究综述:标准、模型与工具

安全攸关软件的可信性关乎生命安全和财产保全,因此,分析评价软件可信性是否符合用户的预期(即软件可信评估)至关重要.软件可信评估从主观和客观两个方面度量软件的质量,对软件生产和应用有着重要的意义.综述了可信评估管理中涉及到的标准、模型和工具,而非关注软件度量本身.首先分析对比了软件可信性、可信评估的定义,并在研究了与可信性密切相关的软件质量的联系与区别之后,从相关国际标准、评估涉及的模型(包括质量属性模型、证据模型、分级规范等)以及软件工具支持等方面综述了软件可信评估研究工作.并且区分了这些方面中领域相关、领域无关的不同之处.目前软件可信评估已取得了一定的理论成果,并开发了若干工具辅助进行可信评估,但仍需在通用性、可伸缩性等方面有所加强.

Survey on Software Trustworthiness Evaluation:Standards, Models and Tools

The failure of safety-critical software could result in death, injury, damage to people or loss of equipment or property. Therefore, it is important to evaluate whether software trustworthiness fulfills the user needs (i.e., trustworthiness evaluation). In this paper, we first compare the definition of software trustworthiness and its evaluation. Then, we survey on the software trustworthiness evaluation from different aspects: standards, models, and CASE tools. We study these aspects from the view of domain-independent as well as domain-dependent. In summary, there is great progress being made for software trustworthiness evaluation theoretically and practically; still we need to improve its universality and scalability.