Syntactic accidents in program analysis: On the impact of the CPS transformation

We show that a non-duplicating CPS transformation has no effect on control- flow analysis and that it has a positive effect on binding-time analysis: a monovariant control-how analysis yields equivalent results on a direct-sty le program and on its CPS counterpart, and a monovariant binding-time analy sis yields more precise results on a CPS program than on its direct-style c ounterpart. Our proof technique amounts to constructing the continuation-pa ssing style (CPS) counterpart of flow information and of binding times. Our results confirm a folklore theorem about binding-time analysis, namely that CPS has a positive effect on binding times. What may be more surprisin g is that this benefit holds even if contexts or continuations are not dupl icated. The present study is symptomatic of an unsettling property of program analy ses: their quality is unpredictably vulnerable to syntactic accidents in so urce programs, i.e., to the way these programs are written. More reliable p rogram analyses require a better understanding of the effect of syntactic c hange.