An ideal random test (Section 5.2) requires a random test sequence allowing testing of all the specified cases and only the specified cases corresponding to the circuit under test. In practice, the generation is based on some algorithm generating a pseudorandom test sequence. Such a sequence has two features: 1) it is not truly random since the generation is algorithmic (from some initial state of the generator, the pseudorandom sequence is always the same); 2) it presents some randomness properties such that it is a good approximation of a truly random one. We shall mainly use the word random to qualify the test sequences. The word pseudorandom will be used in some specific cases.
This chapter presents the needs in Section 10.1 (i.e., the conditions to be satisfied by the random test sequences). Software and hardware generations taking these needs into account are presented in Sections 10.2 and 10.3 respectively.