Anomaly Detection of Functional Programs Based on Global Dataflow Analysis
Mizuhlto Ogawa, Satoshi Ono · Systems and Computers in Japan · 1990
Abstract This paper proposes a new anomaly detection algorithm based on the global dataflow analysis, considering the first‐order functional program. The anomaly detection is made concerning two points: the irrelevant object which does not affect the result by evaluation; and the diverged object which will fall in an infinite loop without being terminated if an evaluation is made. The detection algorithm is based on the global functional analysis technique which determines the computation path arriving at the result of computation. Because of this property, the algorithm has a higher detection ability than the cross‐reference analysis, which simply traces the syntactic definition and reference relation. Using the proposed algorithm, the bug which may diverge the computation is detected effectively automatically, and is often produced in the program using a stream which is conceptually an infinite list.