ABSTRACT
This chapter presents generalizations of data flow frameworks based on mathematical abstractions and provides lattice theoretic modelling of data flow frameworks. It discusses some properties of Complexity for General (CFG) that is relevant to round-robin iterative data flow analysis. The data flow variables are related through equations which are then solved to get data flow values at the program points. The transformations effected by basic blocks on data flow values are called flow functions. The essential properties of flow functions and merge operations are identified as part of the generalization. Systematic computation of data flow values requires that the concept of approximations of data flow values and the operation of merging data flow values should satisfy certain properties. In the context of data flow analysis, the meet operator is used to merge data flow values along different paths and reaching a join node of the underlying CFG.
