Enabling Certification through an Integrated Comprehension Approach

ion, by virtue of producing an abstract model similar to the program model, works in a manner that facilitates bottom-up comprehension. It thus forms a perfect foil for slicing by (automatically) providing program models to the analyst at specified levels of abstraction, that can be used to form the hypotheses upon which the slicing criteria are based upon. Further, the abstracted model can be used as a basis for model checking, to formally verify properties of the system and ascertain the validity of the program, or generate an error trace as a counter-example. This model checking provides the framework for detecting design faults and errors of omission that slicing cannot easily detect. A methodology for combining abstraction with slicing is proposed in [JI04]. Future Directions. To realize this integration, the analyst would need to be provided with an automated tool that combines slicing with model abstraction. The tool would ideally take as input a user-defined criterion and generate a corresponding program slice and abstract model for the same. The criterion for slicing and abstraction could be defined independently, or one could be derived from the other (Bandera [CDH00] and the Promela slicer [MT98] use temporal properties to derive the slicing criterion; a similar method could be developed for the other direction as well). The integration with abstraction could be extended to other slice refining techniques like conditioned and amorphous slicing as well. Using these techniques would result in smaller slices and correspondingly, more refined models; thus further improving the efficiency of the comprehension process.