This article proposes a temporal and parametric specification language (ParTraP) developed for the verification of execution traces. The language extends specification patterns with nested scopes, real-time and first-order quantification over the data inside a JSON trace, while remaining pragmatic. Its design was directed by a case study in the medical field (computer aided surgery). The paper briefly presents the case study and details the design rationale, syntax and semantics of the language. The language has been implemented and several properties have been successfully evaluated over a corpus of 100 surgery traces.

The verification of software intensive medical devices can largely benefit from the analysis of their execution traces. Trace points can easily be added to the software, and traces can be used at several stages of the development and maintenance process. In this paper we focus on the TKA system and identify 15 representative properties that should be fulfilled by its traces. We also identify several stages in the product lifecycle where these properties should be evaluated. These properties put requirements on what should be expressible in a trace property language for medical devices.

No abstract

No abstract