Analysis of concurrent software by cooperative application of static and dynamic techniques

Stand-alone techniques for the analysis and testing of the synchronization structure of concurrent programs have recetly been developed. These techniques are able to detect, for example, task blockage, including deadlock. Static analysis provides firm results, but has limited applicability and is potentially expensive. Dynamic analysis makes fewer assumptions, but its assurances are not as strong. This paper presents strategies whereby the two can be employed jointly to advantage. Dynamic analysis can be used to further investigate results from static analysis, and vice versa. Their joint use can be facilitated by an apporopriate implementation, some principles for which are outlined