This paper has been submitted for consideration for an upcoming Special Issue of the International Journal of Quantum Information with the theme 'Distributed Quantum Computing'.
In due course the reviewers will post their reports into this thread, at which point the author can enter into an exchange with them and/or revise the manuscript.
Third parties are also welcome to contribute to this thread, however the eventual decision of the editors (Dan Browne and Simon Benjamin) will normally be based principally on the reviewer and author postings.

First of all, thank you for your valuable comments. Here is our response, and a summary of the changes we subsequently made to the manuscript.

We have included a summary of the manuscript in the introductory section. The references have been fixed, and now names of all authors are given. We have also added several of the important papers concerning both EIT and STIRAP techniques.

Concerning the bit flip operation in atomic ensembles, we agree that in the case of a matter qubit implemented as an atomic ensemble, the bit flip is not a trivial operation. In fact, for a reliable bit flip operation one has to make use of the dipole-blockade mechanism. In the manuscript, we review the concept of an atomic ensemble as a single qubit system and analyze in detail a scheme for single-qubit operations in atomic ensembles.

Regarding fault tolerance: our entangling protocol is capable of creating cluster states of any degree of connectivity. Since a 3D cluster lattice can be used to efficiently implement planar surface codes, one can exploit the topological error-correction capabilities of the these codes to perform fault-tolerant quantum computation with our entangling procedure.

We also state that the protocol is more efficient than any previously proposed probabilistic scheme with realistic photodetectors and single-photon sources. The justification is the following: every run of our procedure gives an entangled state of two atomic ensembles with success probability , where is the combined detection and source efficiency. The double-heralding protocol produces an entangled state of two matter qubits with success probability . The protocol proposed by Lim et al. requires on average two repetitions to realize the desired gate operation and also relies on multi-photon detection. Our protocol requires only single photon detection. In general, number-resolving photodetectors are not required. However, a reliable photon counting detector with low dark count rate would be able to herald any error in the procedure increasing the fidelity close to unity.

We would like to kindly thank reviewer for this detailed review. We find all your comments valuable and very helpful. Here, we present a summary of the improvements we made to the manuscript based on reviewer’s report.

a. We agree that the effect from the decay processes (occurring with rates of order Hz) cannot be neglected for Rydberg states given in the manuscript ( or ). Higher Rydberg states suffer from decay processes with lower rates of order Hz. However, for this specific experimental implementation it cannot be justified to neglect these effects. Hence, the fidelity of the single-qubit operation will be affected by the decay processes. Nevertheless, the fidelity can still be as high as . We would also like to underline the fact that the error from the decay processes can be neglected in the implementation of our entangling protocol since following successful entanglement preparation the state of matter qubit is quickly stored in the long lived atomic states and .
b. We have added the following explanation of the notation: states and represent two single photons propagating along slightly different paths through upper and lower arm of the interferometer, and interacting with atomic ensembles and , respectively. We agree that although no phase difference appears between two paths (pairs) and until atomic ensembles, the second half of the interferometer after atomic ensembles requires phase stabilization.
c. Considering the ratio of : high probability of absorption requires strongly focused light fields with small area . The improved ratio of and therefore higher probability of absorption for this experimental implementation is owned to the stronger focusing relative to the wavelength of a single-photon pulse. Naturally, the focusing regime is limited by a size of the atomic sample and diffraction limited area of a single-photon pulse. To render the probability of absorption close to unity one may apply a mode converter (shaper) to single-photon fields.
d. On page 31, we state that if one can employ stronger focusing and/or mode converter (shaper) then fidelity of the entangling operation should be rendered close to current fault-tolerant thresholds of the topological codes.

Concerning the typos:
a. The figure 2 caption has been corrected.
b. We believe that indices in Eq. (25) are correct. Eq. (25) follows from the Schrödinger equation and definitions of interaction Hamiltonian (Eq. 17) and state vector of an atomic ensemble (Eq. 21).

The authors have made the appropiate changes I requested and so I am happy to recommend this paper for publication

We would like to thank both reviewers for their time and effort in refereeing our manuscript. Your comments have led to significant improvements.