Theoretical Aspects of Consumer Metrics for Safety & Privacy

Abstract

Software metrics are a matter of academic interest. Consumers, software developers, and other practitioners seem less interested in functional size than, maybe, in quality metrics. Functional size is useful for predicting cost of development; however, with today’s software development techniques and wide availability of open software, its size has become less important. Today, compliance, usability, and connectivity, are the most important cost drivers when developing software-intense products. However, functional size is essential for software testing. This is not yet widely accepted because today’s best practices in testing focus on code coverage. However, software must be tested against functionality. Code is usually not available for cloud functionality such as map services. Software-intense products, for instance when driving autonomous vehicles, or powering robots for daily life, rely on functionality from various origins, not on embedded software alone. This paper sketches the theoretical framework for software metrics that help consumers assessing the level of safety and privacy of their software.