RAPID PROTOTYPING OF DATA FLOW ANALYSIS ALGORITHMS USING SET ORIENTED LANGUAGE (SOL)

摘要

Rapid prototyping of computer algorithms as real programs is important from the viewpoint of cost-performance of software development. In this paper, we show how efficiently intraprocedural data flow analysis algorithms are prototyped as programs using the set oriented language SOL in comparison with some traditional implementation using iterative algorithms. Four important representative problems are pointed out as an important subclass in intraprocedural data flow analysis problems. They are, available expressions, reaching definitions, live variables, and very busy expressions. These problems are very similar in that almost any algorithms to solve one of these problems can be used to solve the other problems. Therefore, in this paper, we focus on "available expressions" to show the effectiveness in rapid prototyping.