The authors wrote an R package to calculate the normalized NODF, a metric that can fairly compare nestedness across networks. The R package makes the application of the normalized NODF possible for complex analyses and large networks. I find the R package timely and easy to use. The paper is also clearly written. Given the importance of nestedness in network ecology, I am sure that the availability of this package would be of interest to a wide audience.

Below I outline some minor issues. I am sure that the authors can address them easily.

Line 68: I would recommend adding more literature on this issue. Clearly, Song et al. 2017 was not the first one to notice this issue. For example, doi: 10.1016/j.ecocom.2015.01.004 also noted that the z-scores of NODF suffer from the correlations. Outside ecology, this issue was long-known in statistics. For example, doi: 10.1198/000313006X152649 and doi: 10.2307/3802789.

Line 85: Why the computation of max(NODF) is NP-hard? I think the authors should elaborate more on this point.

Line 86: Since the computer scientists have not proved that NP != P, I don't think NP-hard implies that NP-hard problems cannot be solved in polynomial times.

Line 99: I fail to see that permutation approaches would suffer from the computational problem. In common null models, the community size and the number of species interactions are fixed, and thus the maximum NODF would be the same for all permutated topologies. I think the authors should find some other problem or elaborate more on this issue.

Eq. 1. Why A[1,5] = 0? I thought it should be 1.

Line 209-210:

Line 266: Add 'a' to 'Nestedness is pervasive pattern in ecological systems.'

Fig 1a & b: The colors are difficult to tell apart all the algorithms. It might be good to also add shapes to distinguish these algorithms. Also, the coloring makes it difficult to link Algorithm Song with Algorithm 0 or Algorithm Simmons old with Algorithm 2.

Fig 1c: It might be good to say max NODF on the label of the x-axis.

Line 242: It would be good to discuss some example whether the sacrifice of quality would be problematic. For example, it might be worth mentioning that doi: 10.1111/1365-2656.12964 found that the corrected maximum nestedness does not affect the qualitative results of Song et al. 2017. Or perform the two applications using different networks (perhaps in Supplementary Information). This would be a more concrete recommendation to users.

Fig 2: The caption of Figure 2a is very brief. Might be good to expand a bit on the implications of Figure 2a (since most readers read the figures more carefully than the text).

Line 319: No journal information.

The authors successfully addressed all of my concerns. I especially appreciate the authors' efforts in proving that it is indeed NP-hard to find the maximum NODF. I think this paper is ready for publication.