Document Type

Technical Report

Publication Date

Winter 1-1-1979


An approach to the analysis of concurrent software is discussed. The approach, called anomaly detection, involves the algorithmic derivation of information concerning potential errors and the subsequent, possibly non-algorithmic determination of whether or not the reported anomalies are actual errors. We give overviews of algorithms for detecting data-usage and synchronization anomalies and discuss how this technique may be integrated within a general software development support system.