In 2.12 minutes, an IBM 7090 program found four input tests (for an 8-input parity check circuit) whose outcome determines whether any one of 102 possible failures occurred. For any single-output combinational circuit, with no more than 35 input variables, the program computes the set of all inputs detecting a given failure — the essential novelty of the method. These sets, one for each failure, are then processed to find a small subset of tests which detect any failure. The underlying method extends to the diagnosis of circuits with feedback.
No takes yet. Share an insight, caveat, or question.
Galey et al. (1964) studied this question.