A verification tool to measure software in critical systems

Abstract

Previously, software metrics have been established to evaluate the software development process throughout the software life cycle, and have been effective in helping to determine how a software design is progressing. These metrics are used to uncover favorable and unfavorable design trends and identify potential problems and deficiencies early in the development process to reduce costly redesign or the delivery of immature error prone software. One area where design metrics plays an important role is in the identification of misunderstandings between the software engineer and the system or user requirements due to incorrect or ambiguous statements of requirements. However, the metrics developed to date do not consider the additional interface to the safety engineer when developing critical systems. Because a software error in a computer controlled critical system can potentially result in death, injury, loss of equipment or property, or environmental harm, a safety metrics set was developed to ensure that the safety requirements are well understood and correctly implemented by the software engineer. This paper presents a safety metrics set that can be used to evaluate the maturity of hazard analysis processes and its interaction with the software development process.