A new parallel processing language and a scheduling algorithm for parallel processing systems are introduced. The design objectives of the language are clarity, understandability and modifiability. Considering these objectives, we give frameworks of the language. In this language, an algorithm and its constraints are described separately, and hierarchical structure of a problem decomposition is described explicitly. We also give a scheduling algorithm in which processing states are considered. As the state change incurs overheads, the scheduler makes trade-off between wating time of tasks and total loss. Finally, we describe the outline of the scheduler in this language for comprehension of both the language and the scheduler.