The next two faults categories, excluding the OMISSION faults, emulate specific programming errors common to kernel code according to earlier studies [Sullivan and Chillarege, 1991; Christmansson and Chillarege, 1996].
The second category of fault changes individual instructions in the text segment. These faults are intended to approximate the assembly-level manifestation of real C-level programming