Assisting potentially-repetitive small-scale changes via semi-automated heuristic search

Abstract

When a software system undergoes modification, a given change might need to be repeated throughout the codebase. While the change itself may not be difficult to implement, discovering other locations where this change should be applied (if any exist) can be onerous. Syntactic differences in otherwise semantically similar code can render traditional search techniques ineffective. This paper describes a heuristic search technique to help find the locations required to complete a repetitive small-scale change (RSC). By observing the developer perform a change once, it is possible to infer semantic information about that change and to automatically suggest locations where the same change ought to be made