I am reading Harris and Harris's Digital Logic and Computer Architecture. At the outset, let me define (precisely as they and, as far as I know, everyone else does) a combinational logic circuit (I emphasize that we are at the logic circuit level of abstraction) as a circuit whose outputs depend only on the current values of the inputs.
In the text, they give the rules of combinational composition (RCC) as:
- Every circuit element is itself combinational.
- Every node of the circuit is either designated as an input to the circuitor connects to exactly one output terminal of a circuit element.
- The circuit contains no cyclic paths: every path through the circuit visits each circuit node at most once.
They further remark that these rules give a sufficient but not necessary condition for a given circuit to be combinational, which is plainly true.
Now they give the following circuit as an example of a circuit which is not combinational, which seems very strange to me:
It is certainly true that this circuit violates requirement 2 of the RCC, but it still seems to me that this circuit is indeed combinational. Thus, I am led to the following two questions:
A) Am I correct that the circuit is combinational and that the authors should have said "the circuit does not obey the RCC but is nevertheless combinational"? Edit: Now that I look at it, I see the trouble is that two gates are driving the same node and so we can potentially have some contention. I guess the point is that our output is not well-determined (i.e. not fixed) by the inputs in this case.
B) If I am right about A) (Per the edit above, I am not, but my question here still stands) then I am immediately led to wonder why on Earth the RCC are useful if they are so undiscerning as to not "recognize" the circuit above (or, rather, some other combinational circuit which does not obey RCC) as combinational?
