An optimization for information flow analysis

Various optimizations have been suggested for reducing formal flow violations produced by a standard information flow security analysis for the certification of programs or specifications of computer systems. These optimizations are specific to particular source language constructions and rely on the analysis tool recognizing these constructions through syntactic processing in order that the optimization(s) might be applied. A new optimization for eliminating formal flow violations, suggested by the semantics of the text under analysis, is presented. It is shown that the syntactic optimizations mentioned above can be viewed as special cases of the generalized optimization.<>