A batch process is an RT program very much like the one that handles interactive communication with the SINTRAN III system from terminals. The task of a batch process is to execute jobs on a batch-input-file, one at a time.
In principle, there may be an unlimited number of batch processes running parallel to each other and other activities in the system. The maximum number of batch processes is determined at system generation time.
The extra overhead introduced by a batch process is approximately the same as by adding an extra terminal. A batch process is initiated by the @BATCH command, and terminated by the @ABORT-BATCH command. Both commands are issued only by the user SYSTEM. A batch process will be in one of the following states
- this means that the batch process has not been started.
- this means that the batch process has entered waiting state because the batch queue is empty.
- this means that the batch process is working on a job.
When appending a batch-input-file to a batch queue in idle state, it will be activated. When end-of-file is reached on a batch-input-file, the batch process fetches the next batch-input / output pair of the queue, and continues taking input from the new batch-input-file and giving output to the new batch-output-file. When the batch queue for this process is empty, the process enters idle state again. 
- Norsk Data Document ND–60.050.06 SINTRAN III USERS GUIDE , section 3.7.2