Efficient scheduling of a parallel program represented by a Directed Acyclic Graph (DAG) onto a distributed system involves a tradeoff between the schedule length and inter-processor communications. In this paper we present an efficient scheduling algorithm that builds upon our preliminary Recursive Critical Path Approach (RCPA) resulting in significant improvements in performance metrics (i.e., processor requirement, inter-processor communication, and schedule length). Extensive simulations on random DAGs show that 35% of the clusters generated by our approach are inter-processor communication free as opposed to 0% of a top leading advanced technique (i.e., the LG Algorithm). Furthermore, the average number of processors used is 27% less than that of LG, while the schedule lengths are 5% better than LG and 21% better than our former approach.