A theory of conflict resolution in AI planning is presented. The paper proposes a strategy for handling conflicts in a global manner, using constraint satisfaction techniques. It shows that this strategy can reduce the search space. Yang states that in planning problems with a large number of conflicts, the benefits from the reduction of the search space outweigh the extra overhead incurred by the strategy. His empirical results support this claim, but unfortunately his comparisons are based on TWEAK, a planner known for its inefficiency. Nonetheless, his arguments are convincing. Overall, this paper is well written. The intended audience is people interested in the classical planning paradigm. Yang provides a concise description of the preliminaries, but it is a bit dense. Thus, some familiarity with traditional planning research would be helpful.