Allegations as aids to static program testing

Abstract

In static program analysis, a program is examined without execution in an attempt to anticipate possible sources of error. Possible errors detected in this way can rarely be considered certain to occur because of the impossibility of infallibly determining the executability of a given program path. Current heuristic systems for making this determination are costly and uncertain. Hence the use of allegations—user supplied unverifiable statements designed to provide answers to executability questions—is suggested.