Dear Peter,

many thanks for the opencharm paper, it is very well-written to
present a vast number of important results while staying concise!

Below are my comments to paper draft v5.1 (apologies in advance for the
lengthy mail and for not having carefully read the previous version).

1) p.1, 1st para, line 7: citation formatting
 [4]-[9]  --> [4-9]
2) p.1, end of 1st para. of subsection 2: citation formatting
 [16]-[22] --> [16-22]

--> done; it took us sometime to figure out how to do it...


3) p.2, 1st para:
 3a) to me the word "typically" looks a bit confusing, i thought it was
     just exactly 300 mkm (cathode+anode) everywhere in ITR
 "__typically__ 300 um pitch" --> "$\sim 300~\mu\mbox{m}$ readout pitch"

--> OK

 3b) there are too many long phrases in parentheses in this paragraph, i'd
 re-shuffle the parentheses e.g. as follows:
 --> ... (ITR) --- microstrip gas chambers with gas electron multiplication
     and $\sim 300~\mu\mbox{m}$ readout pitch~\cite{ITR}, and a coarse
     grained outer tracker (OTR) --- honeycomb drift cells with 5 mm and
     10 mm cell diameters~\cite{OTR}. The momentum resolution can be
     parameterized as $\sigma_p/p = 1.61+(0.0051\cdot p~[\mbox{GeV}/c])~\%$
     \cite{OTR}, where $p$ is the particle momentum.

--> done

 3c) formatting of the momentum resolution "0.0051pc/GeV)\%" looks awkward,
     better formatting similar to that of the OTR paper:
 --> $\sigma_p/p = 1.61+(0.0051\cdot p~[\mbox{GeV}/c])~\%$

--> I have some preference for the original version, but will adopt the
--> one from the OTR paper.

 3d) i'd combine the two references to two OTR papers [18] (II) and [19] (III)
 into one reference and also add their paper number (I) -- see comment (32j)
 to the list of citations below.
 (Note, that sigma_p/p refers to the III OTR paper, though probably it's ok
  to cite the same paper in two consecutive sentences as in (3b) ...)

--> done

4) p.2, 3rd paragraph:
 I think that the complete 3rd paragraph describing ECAL and MUON can be
 safely dropped out since these subdetectors were not used in the described
 analysis. Just move corresponding citations of the corresponding detector
 papers to the 2nd paragraph on page 2.

--> I left it as it was for the sake of completeness.

5) p.2, line 3 of 4th para.:
 the abbreviation (VDS) has been already defined on p.1, leave either
 --> "... following components: the vertex detector, the OTR ...", or
 --> "... following components: the VDS, the OTR ..."

--> done


6) p.2, Table 1:
 the errors on luminosity numbers are missing. Though, in this context
 that may be ok. The dominant lumi error is the systematical one which
 is mentioned separately in Table 5 (see the comment (18) on lumi errors
 to Table 5).

--> I think this should remain as it is. We do not want to make this table 
--> too busy. We also refer to the lumi paper.

7) p.3, 5th para (left column, line 5 from the bottom of the page):
 I just wonder whether one needs to mention that $x_F$ is "the Feynman x",
 or not. At least, in case one needs to, then here is the place where it
 is first used in "$-0.15 < x_F < 0.005$".

--> done; we use is it already in the abstract...

8) p.4, Table 2:
 8a) D+ -> K-pi+pi+:
     cut value "L_K(K) > 0.5" differs both from the analysis Note, Tab 10,
     and Samo's thesis (Tab 4.2), where it is L_K(K) > 0.33.

     (The same cut "L_K(K) > 0.5" for D+ is also mentioned in the
      text on p.3, line 8 of the 2nd para of subsection 3.1)
 8b) D+ -> K-pi+pi+:
     cut values "b(K)b(pi_1)b(pi_2) > 106" and
     "sqrt[3]{b(K)b(pi_1)b(pi_2)} > 4(t-t_0), \;\; t_0=2.48"
     differ from the analysis Note, Tab 10 and Samo's thesis (Tab 4.2).

--> Indeed, some cuts got a final touch by Bernhard to make the two analyses
--> easier to compare. The final set of cuts is better recorded in his
--> note.

9) p.6, 1st para, line 3:   "... to the parameterizations"
   in some places you write "parameterizations" and in others "parametrizations"
   afaik, both are correct spellings, but better use one (please search
   through the text in the editor for other occurrences of these two forms).

--> done

10) p.6, Eq. 1:
  10a) probably needs a colon before the formula
  10b) and needs a comma after the formula

--> I think that a comma after 2 is just right.

11) p.6, line 12:
  "and $x_b$ was assumed to be" (the assumed value missing)

--> a % sneeked in ....

12) p.6, last paragraph of subsection 4:
  "... the same reconstruction codes as that ..."
  --> "... the same reconstruction codes as those ..."

--> done

13) p.6, Eq. 3:
  "\Delta\sigma_{pA}" --> "\Delta\sigma_{pA,\,i}"

--> done

14) p.6:
 14a) add dots after the Eqs. 5, 7, 8.
 14b) add colons before Eqs. 6, 8.

--> done

15) p.6, right column, line ~10: probably the "A" should be in italics
 "..., a linear A dependence ...":
 --> "..., a linear $A$ dependence ..."

--> done

16) p.6, right column, lines 10-12: a verb seems to be missing
 "of the production cross section assumed, ..."
 --> "of the production cross section was assumed, ..."

--> done

17) p.6, right column, end of the next-to-last paragraph:
 I think one should add a comment on how the estimates of the other two
 systematic errors (track reco and particle identification) were obtained.
 This is especially important for the PID systematics since the whole
 analysis heavily relies on the RICH PID.
 I would add something like:
 --> "The systematic errors of track reconstruction and RICH particle
      identification (1.5\% and 2.0% per track correspondingly) were
      estimated using decays of $K0_S$ and $\phi$ as sources of pions
      and kaons~\cite{Pernack,Kupper}."

 Strictly speaking, in the Note and Samo's thesis it's mentioned that a momentum-
 dependent efficiency correction factor determined from the above PID studies was
 applied on top of the MC efficiency. I don't know if that deserves to be mentioned
 in the paper, or if that would already be too much of a technical detail...

--> I adopted your suggesting, without citing the two reports. I think that 
--> the calibration method we employed is quite standard.

18) p.7, Table 5.: Luminosity systematical errors.
 The lumi errors quoted in Table 5 (3.4%) are the same as in the analysis
 Note subsection 6.1.3 which seems to be out of date. In Samo's thesis,
 subsection 7.1.3 the lumi errors correspond to the ones of the lumi paper
 submitted for publication.
 The difference between the two is probably negligible, but please check
 which lumi errors were actually used for the results in the paper.

--> Checked by Samo. It is 3.7%, corrected. 

19) p.7, 3rd paragraph of subsection 5.2:
 To my view the procedure of simultaneous fit is too complex and there is
 no or only a little use in describing it without giving much more details
 (e.g. other background parametrization formulas which were tested,
  motivations for using the same functional form of background for
  D0 and D+ (e.g. though in mass the D0 and D+ background differ so much),
  mentioning that for D*+ not a side-band, but wrong-sign combinations
  from the signal window, etc.). Note also, that the fit procedure as
  described now is much more detailed for p_T than for x_F.

 So I would propose to rephrase the 3rd paragraph giving a reference to
 Samo's thesis for a more detailed description of the fitting procedure.
 One can also combine the resulting phrase into the 2nd paragraph:

e.g.
 --> "The likelihood included the paramet?rization given in Eq. 1
      to describe the signal events, with $\langle p_T\rangle$ and $\beta$
      being common to all $D$ meson species, and a three bell-shaped
      terms (with their functional form tested in the sideband regions)
      to account for kinematical distribution of the background
      separately in each $D$ meson signal channel~\cite{Kupper}."
or
 --> "The likelihood included the paramet?rization given in Eq. 1
      to describe the signal events, and three bell-shaped terms
      to account separately for kinematical distribution of the
      background in each $D$ meson channel. The parameters
      $\langle p_T\rangle$ and $\beta$ were common to all the $D$ meson species.
      The functional forms used to describe the backgrounds had been tested
      in the sideband regions and both their normalization and shape
      parameters were allowed to vary independently for each of the $D$
      meson signal channels~\cite{Kupper}."


--> I do not think we should go into a lot of explanations here, so I would
--> prefer to keep it as it is


20) p.8, left column, line 11:
 I think, if the value <p_T> = 0.9~Gev/c extracted from Pythia MC has any
 important meaning, then one should mention the version of Pythia (isn't
 it the default hera-b Pythia 5.7?).

--> We mention it in section Eff. determination.

21) p.8, right column, last line:
 concerning the sigma(D+)/sigma(D0) predicted by Pythia: same as (20), by
 which version of Pythia, was it additionally tuned, etc.?

--> I am afraid nobody really remembers. I talked about it with Thomas Lohse,
--> and he was only sure abot 5.7. Neither does Achim. I therefore assume we 
--> have not made any attempt to tune it.

22) p.9, left column, line 7: since [5] is NA27, and [9] refers to E769
 "... results of E769 and NA27 [5,9]."
 --> "... results of NA27 and E769 [5,9]."

--> done

23) p.9, left column, line 9:
 "... ways ([1] and references therein) ..."
 --> "... ways (see [1] and references therein) ..."

--> done

24) p.9, right column, line 2:
 24a) introduce some symbol for the asymmetry, e.g. \mathcal{A}(D) which
      can be also used in the header of Table 10
 24b) use an \mbox{} in the subscript of non-leading \sigma or write
      \sigma{non~LP} to avoid that "non" and "LP" fuse together, e.g.:
 --> "$\mathcal{A} \equiv (\sigma_{\mbox{\scriptsize LP}}
       -\sigma_{\mbox{\scriptsize non LP}})/(\sigma_{\mbox{\scriptsize LP}}
       +\sigma_{\mbox{\scriptsize non LP}})$"

--> done

25) p.9, right column, lines 10-11:
 In the erratum to the paper of E769 [9], it is said that the errors on
 the asymmetries should be increased by factor 2 compared to their paper,
 --> correct the errors of the E769 asymmetries to:
     0.06+-0.13, 0.18+-0.11, and 0.36+-0.26.

--> Thanks, we missed this!

26) p.9, Table 10 + right column, lines 10-12:
 Listing of all the values of asymmetries measured by other experiments in
 the text makes it more difficult to read. Maybe one could collect them in
 Table 10, which then also could be referred to as a "comparison" with
 other experiments? E.g.:
 -->
 \begin{table}[h]
 \caption{Leading to non-leading particle asymmetries $\mathcal{A}$
 (defined in the text) in the visible range, compared with asymmetries
  published by other experiments. The visible range of E769 is $x_F > 0$,
  and that of E789 is $0 < x_F < 0.08$. \label{label-of-tab-10} }
 \begin{tabular}{lccc} \hline
    & $\mathcal{A}(D0)$ & $\mathcal{A}(D^+)$ & $\mathcal{A}(D^{*+})$ \\ \hline
  HERA-B & $0.10\pm0.09\pm0.05$ & $0.14\pm0.11\pm0.06$ & $0.23\pm0.17\pm0.06$ \\
  E769~\cite{E769} & $0.06\pm0.13$ & $0.18\pm0.11$ & $0.36\pm0.26$ \\
  E789~\cite{E789} & $0.02\pm0.06$ &  ---          & --- \\
 \hline
 \end{tabular}
 \end{table}

--> Here I disagree. I do not think we should put too much emphasis on 
--> this topic, and such a table does it.


27) p.10, Table 11
 27a) 1st line of the caption of the Table:
      "dependency" --> "dependence"
--> dependency is in principle OK as well, but also has other meanings -> done

 27b) 1st line of the caption of the Table:
      "... parameter $\alpha$ and a"
      --> "... parameter $\alpha$ and its" (?)
 27c) try if centering of the "$\alpha$" in the header
      of Table 10 would make look  nicer (?)

--> done


28) p.10, left column, line 2: please clarify which version of the lumi
 errors was used. (e.g. in Samo's thesis the lumi numbers correspond to
 the lumi paper and lead to the typical value of the error on alpha ~2.2%).

--> The lumi error corrected to 2.2%


29) p.10, left column, line 8 of the Summary:
 "... were measure: ..."
 --> "... were measured: ..."

--> done

30) p.10, 2nd and 3rd paragraph of the Summary:
 To my view the summary looks too bulky because of so many
 values of all the main results. Wouldn't it look nicer just to
 refer to Tables 6 and 9 (and perhaps also collect the two P_V
 measurements in Table 9)?

--> I always understood that the summary should cover most of  
--> the results of the paper, so I would prefer to keep it as it is now.


31) p.10, left column, line 8 from bottom: add citation to the PDG:
 "... from $e^+e^-$ collisions."
 --> "... from $e^+e^-$ collisions~\cite{PDG}."

--> Skipped comparison to e+e- (see my reply to Wolfgang).

32) pp.10,11 (References)
 32a) many references lack a comma between the author list and the journal name
 32b) all references lack a dot in {\it et al.}
 32c) [1] "H.~W\"ohri" --> "H.K.~W\"ohri"
 32d) [2] "G.~Ridolfi G" --> "G.~Ridolfi"
 32e) [3] NPB327, (1989) 49: there also exists an erratum (some formulas
      for differential cross section in the limit of high sqrt{s}, p_T >> m_c)
      --> add "(Erratum-ibid. {\bf B335}, (1990) 260)"
 32f) [9] PRL77, (1996) 2388: formatting of the reference to the erratum
      "(Erratum-ibid. 1998 {\bf 81} 1537)"
      --> "(Erratum-ibid. {\bf 81}, (1998) 1537.)"
 32g) [10] "P.~Nason P" --> "P.~Nason"
 32h) [11-14]
      "I.~Abt (HERA-B Collaboration) {\it et al}"
      --> "I.~Abt {\it et al.} (HERA-B Collaboration),"
 32i) [14]:
     * extra "2007" after "et al"
     * "(2006)" --> "(2007)",
     * "Eur. J. Phys. C" --> "Eur. Phys. J. C";
     in total, the Psi' reference should look like:
     -->"I.~Abt {\it et al.} (HERA-B Collaboration),
          {\it Eur. Phys. J. C} {\bf 49}, (2007) 545."
 32j) [18,19]: (OTR paper, parts II and III)
     * [19]: "NIMA576, (2007) 313" --> "NIMA576, (2007) 312"
     * I'd propose to combine the OTR papers under one \bibitem and add
       there also the 1st part of the OTR trilogy. By the way, yesterday I
       found that SPIRES knows only about its e-print version,
       therefore I think it would be good to cite the NIM versions of
       the complete OTR trilogy.
     --> \bibitem{OTR} H.~Albrecht {\em et al.,} {\em Nucl. Instrum.
           Methods A} {\bf 555}, (2005) 310;
           ibid. {\em A} {\bf 541}, (2005) 610;
           ibid. {\em A} {\bf 576}, (2007) 312.
 32k) [20]: "I.~Ari\~no I" --> "I.~Ari\~no"
 32l) [23]: "R.~Battiston R" --> "R.~Battiston"
 32m) [24]:
     * lumi paper is a collaboration one (thus in SPIRES and in the arXiv)
     * [hep-ex] --> [physics.data-an]
     "M.~Bruschi {\em et al} arXiv:0706.0131 [hep-ex] (2007)"
     --> "I.~Abt {\em et al.} (HERA-B Collaboration), arXiv:0706.0131
          [physics.data-an] (2007)"
 32n) [25] "W.M.~Yao" --> "W.-M.~Yao"
 32o) [26] "T.~Sjostrand" --> "T.~Sj\"ostrand"
      (well, though SPIRES knows him as T.~Sjostrand without umlaut...)
 32p) add citations to Robert Pernack's and Samo Kupper's theses
     (cf. comments (17, 19)):
\bibitem{Pernack} R. Pernack, PhD thesis, University of Rostock, Germany, (2004).
\bibitem{Kupper} S. Kupper, PhD thesis, University of Ljubljana, Slovenia, (2007).

 32q) probably remove ref. [30] E769, PRL77, (1996) 2392
     (cf. comment (19) -- it seems that this reference is needed only for
      the 3rd paragraph of subsection 5.2 for which I propose a rephrasing.)

--> all done (except 32p+q)

With best regards,
Dima.