The safety-critical software evaluation assistant (SEA)

The Computer Safety and Reliability Group at Lawrence Livermore National Laboratory (LLNL) is researching the evaluation of software used in safety-critical applications. This paper describes one, of the research and development efforts currently underway to model the software evaluation process and to develop a software evaluation tool. One of the primary techniques available for determining the safety of software proposed for use in safety-critical applications is to evaluate the software development process and the resulting products. This model of the evaluation process was influenced by several factors the underlying motivation was to identify, control and reduce the risk inherent in building safety-critical software systems. This prototype tool, the Software Evaluation Assistant (SEA), assists and guides evaluators as they analyze safety-critical software. SEA describes specific evaluation goals, provides a brief overview of the specific evaluation process, identifies potential, risks of not performing the evaluation, identifies the skills required to carry out the evaluation of a particular topic, identifies the material that should typically be available for the evaluation, and poses questions used to examine and rate the software item