It is a common belief among engineers that symmetric operations are better for combinational logic synthesis. The authors here suggest that asymmetrical operators {<, =>} have better expressiveness, based on experimental results using an exact algorithm.

Two metrics of expressiveness are described in the paper. The first one uses the total gates required to express all 16 Boolean operations on two variables, where the set {<, =>} outperforms traditional favorites NAND/NOR. It is not clear if this metric can be generalized for the case of n variables, and whether asymmetrical operators continue to be better when the number of variables increases.

The second metric takes a given number of gates and measures the number of functions expressible by them. Results presented show that libraries with a mixture of symmetric and asymmetric operations like {<, =} and {=>, ^} do better, and the conjecture on circuits built exclusively with asymmetrical operators is not obvious.

NAND and NOR seem to be better for implementation, as reducing the transistor counts for {<, =>} requires pass transistor logic. More extensive validation of the models presented is required to determine if asymmetrical operators will become viable.

The conclusions reached here are interesting, but the chosen metrics need better justification. The ideas presented here show promising directions for future research.