Hi Tjeerd, 

Two main issues from my side:

1) The comparison between experiment and theory needs to be done with care,
since the experimental branching ratio is time-integrated, while the theoretical
branching ratio is defined from the amplitude level at t=0. 
The translation between the two values is given in Eq.(12) in [arXiv:1204.1735]. 

We suggest to add this sentence after L.17:
"When comparing the experimental and theoretical branching ratios, special care 
has to be taken due to the lifetime difference of the $B_s^0$ system. The correction
can be as large as 9\%, depending on the polarization state and the value of 
$A_{\Delta\Gamma}$ as described in Ref.[arXiv:1204.1735]."

2) Fig.2: The amount of partially reconstructed B+0->hhgamma is very different
between the two final states. Do you understand that? This directly affects the
combinatorial background, which in turn affects the signal yield and thus
the measured ACP.

Cheers, Niels

PS. Minor textual stuff in the attachemnt.


Abstract:
-----------
* [remove the world average number; at first glance it looks like a measurement]

1 Intro
--------
L.3   through b->sgamma one-loop electromagnetic penguin transitions ->
      through the electromagnetic penguin transitions b->sgamma, at one-loop level.
L.23 "could be enhanced up to -15\% [8]." -> "could be enhanced [8]."
L.25 [superscript '2' looks like 'power of 2'. Also in L.114. Move footnote to
     stick it with a word, instead of a symbol.]
L.27 "A measurement of the direct CP asymmetry of the decay B0->K*gamma is also 
     presented."

2 LHCb
------
L.43 on hardware -> in hardware
L.48 low enough so tracks can be -> low enough for tracks to be 
L.51/2 rate to a point ... be recorded.  -> 
     rate such that the HLT2 can perform full event reconstruction to further reduce the data rate.

3 Selection
-----------
L.79 [Rephrase "to reject kaon misidentification". In fact, why not skip the 
full sentence? The fact that charged tracks are identified is already mentioned 
in L.77.]

L.80 (Pions) -> (pions)
L.90 ["for decays in this paper" is obvious. How about:]
"The invariant mass resolution of the selected B candidates amounts to
~ 100 MeV/c2."

4 Fit
-----
L.107 on both side -> on either side
L.110 [The use of 'candidates' is a bit misleading when talking about background, skip?]
       "The mass distribution has been found to be compatible with an exponential ... "
L.111 ['each' suggests many, but you only discuss two distributions; how about:] 
      "with different decay constant for the two channels."

L.118 'or by estimating from the signal yield' ->
      'or by directly estimating the signal yield from the observed signal (Lb->L*gamma)."


L.123 [add charge to B0? If you refere generically to B+ or B0, then the charges on the h+h- 
are not general enough? How about:] "B->h+h-gamma" -> "B+/0 -> hhgamma"  ?

L.125 " radiative decays B+->K*0pigamma " -> " radiative decays of charged B mesons, B+->K*0pigamma , "

L.129 [Phrase more explicit: ] 
The neutral partner decays of those charged B decays -> 
The partially reconstructed neutral B meson decays 

L.133 [Perhaps use symbol 'V" instead of 'K*0(phi)'. Also in Table 1. ]

L.139 'The acceptance is modelled' -> 'The inefficiency at the edges of the mass window is modelled'

5 BR
----
Fig.1: [last sentence of caption: is that sentence needed? Is the residual different from what the
reader naively expects? If not, perhaps it is better to remove the sentence to prevent confusion.]


L.202 [Add that the trigger efficiency is obtained from sim]


6 ACP
-----
L.233 [invert (K+pi-) to reflect the double mis-id better:] 
"(K-pi+)->(K+pi-)" -> "(K-pi+)->(pi-K+)"

L.246 Reference [24] is published; arXiv:1202.6251.

L.251 [specify you use the B0 lifetime?] 
"... determined from data using the known B0 lifetime, by means of the splot technique..."

L.253/4 "by the magnetic field that deflects oppositely charged tracks to different regions in the detector"


7 Results
---------
L.267 'In 1.0 fb-1 of pp collissions collected with' -> 
'With the data sample corresponding to an integrated luminosity of 1.0 fb-1, collected with'


Ref
----
[8] Not published; is this the best reference for NP ACP ?