<div dir="ltr">Dear Michel and dear all,<div><br></div><div>we modified the paper accordingly to our agreement. You can find it at the link:</div><div><br></div><div><a href="http://www.lnf.infn.it/~pisanos/dvcs/prd/prd_dvcs_eg1dvcs_author_check.pdf">http://www.lnf.infn.it/~pisanos/dvcs/prd/prd_dvcs_eg1dvcs_author_check.pdf</a><br></div><div><br></div><div>Let us know if we can now proceed to the final author check.</div><div><br></div><div>Best regards,</div><div><br></div><div>silvia</div><div><br></div><div><br></div></div><div class="gmail_extra"><br><div class="gmail_quote">On 13 January 2015 at 17:30, GUIDAL Michel (57321) <span dir="ltr"><<a href="mailto:guidal@ipno.in2p3.fr" target="_blank">guidal@ipno.in2p3.fr</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">I could agree to something like that. My intention is not<br>
to "erase" the CFF notation that Michel advocates. But my intention<br>
is certainly to keep alive the notation that I advocate (especially<br>
when it is my work which is quoted !).<br>
Just for completeness/addition, I am about to release,<br>
with my (former) grad. student Marie, a paper<br>
on the simultaneous fit of DVCS and TCS (pseudo-)data.<br>
And in TCS, what one fits is the _conjugate_ of the CFF<br>
to the current Eq.8, which would mean that you don't have the same<br>
CFFs, according to the current eq.8, between DVCS and TCS.<br>
With the definition that I advocate (and that I will continue to use<br>
in any case), I can certainly speak of the "same" CFFs: this is<br>
flexible, convenient and esthetic.<br>
Amities,<br>
<br>
Michel<div class="HOEnZb"><div class="h5"><br>
<br>
On Tue, 13 Jan 2015, Silvia Niccolai wrote:<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Dear all,<br>
I have a proposal to solve this perennial issue of the definition of the CFFs.<br>
We could keep in the introduction the CFF definition that is in the current version, but in the section devoted to the CFFs extraction we could specify the definition adopted in the fit-based extraction procedure by Michel et al.<br>
This way all would be consistent and clearly stated.<br>
What do you guys think?<br>
Best regards<br>
Silvia<br>
<br>
<br>
Sent from my iPhone<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
On 13 Jan 2015, at 16:17, GUIDAL Michel (57321) <<a href="mailto:guidal@ipno.in2p3.fr" target="_blank">guidal@ipno.in2p3.fr</a>> wrote:<br>
<br>
Dear all,<br>
<br>
I went quickly through the latest vesion of the article:<br>
<br>
<a href="http://www.lnf.infn.it/~pisanos/dvcs/prd/prd_dvcs_eg1dvcs_author_check.pdf" target="_blank">http://www.lnf.infn.it/~<u></u>pisanos/dvcs/prd/prd_dvcs_<u></u>eg1dvcs_author_check.pdf</a><br>
<br>
I notice that, under the comment/pressure of only one person, the CFF<br>
definition has been changed (Eqs.9 and 10).<br>
<br>
1/I think that the least that should be done is to mention that<br>
some people use other definitions, especially when this is<br>
the work of these particular people which is particularly used in the paper. I remind that the latest review on the field (Rept.Prog.Phys. 76 <a href="tel:%282013%29%20066202" value="+12013066202" target="_blank">(2013) 066202</a>, M. Guidal, M. Vanderhaeghen and H. Moutarde, i.e. signed<br>
by 3 authors which, I dare to think, are not beginners in the field) use<br>
the particular definition that was originally used in this paper.<br>
<br>
2/this absolutely useless and unnecessary change leads now<br>
to inconsistencies in the text:<br>
<br>
l.1172:"In recent years, various groups have developed and applied<br>
different procedures to extract Compton Form Factors<br>
from DVCS observables". With the current definition of CFFs, this<br>
is not true, nobody has tried to extract: "Int(GPD(x,+/-xi,t)/(x+/-xi+<u></u>ieps)dx".<br>
<br>
l.1176:"the CFFs are almost-free parameters". With the current<br>
definition, this is not true.<br>
<br>
l.1187:"to extract the Compton Form Factors". Same as above.<br>
<br>
l.1197:"Given that the size of the error bars reflects the sensitivity of the combination of observables to each CFF". With the current definition<br>
of CFF, this is not true because the different observables have<br>
different sensitivities to the imaginary and real parts of CFFs<br>
(current definitoin ;-) ).<br>
<br>
l.1214:"both sets of CFFs are compatible". This is not correct<br>
with the cirrent defintin of CFFs.<br>
<br>
And the vertical axis of Fig.25 is very ambiguous. Please, let's<br>
avoid the horrible: "-Im(Htilde)/PI..."<br>
<br>
And I have probably forgotten some other instances.<br>
<br>
3/there is absolutely no doubt in my mind that the original notation<br>
used in this paper was the most practical, convenient and esthetic<br>
and I don't understand why one person managed to change the notation.<br>
I certainly recommend to go back to the original one. Especially that the next paper to come in the subject, the DVCS x-sections by HS Jo et al., will use that paricular notation.<br>
<br>
Amities,<br>
<br>
Michel<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
On Tue, 13 Jan 2015, Silvia Pisano wrote:<br>
<br>
Dear Michel,<br>
thanks a lot for your useful comments. We implemented most of them (and did<br>
not report about the accepted ones, since we agree on them).<br>
Below, you can find our answers on the points that need a more detailed<br>
discussion.<br>
Best regards,<br>
Angela, Erin, Silvias<br>
------------------------------<u></u>-----------------------------<br>
- "Eqs 1 and 2 could (should?) be in line within the text, and there is a<br>
minus sign missing in Eq 1."<br>
Thanks for noting the sign, you are right! As to the positioning of the<br>
equations, we preferred to leave them as they are, to give the right<br>
resonance to the quantities introduced. We also slightly modified this part<br>
following some suggestions from Volker.<br>
- "L73-74: not quite correct. “The transverse part of t is the squared<br>
Fourier conjugate of the transverse ….”.<br>
Yes, you are right. We corrected it.<br>
- "Eq8: to be consistent with Eq 6, is not there a factor –pi missing ?"<br>
Yes, we included it.<br>
- "L106: “With a finite number of measurements, a model input is necessary<br>
to deconvolute…” (meaning that with an infinite number of integrals along x,<br>
there are, or could be, mathematical ways to get the x-dependence; this may<br>
be compared to medical imaging reconstituting 3D images from 2D<br>
measurements)"<br>
We preferred to leave it as it is, since we will always have a finite number<br>
of measurements only, and don't wanna be misleading with too many details.<br>
"L135: HERMES statistics on TSA are comparable to exploratory CLAS (see Fig.<br>
22). I would write: “….was also performed. The statistics obtained in<br>
measurements [13,14] do not allow for a 4-dimensional binning of the data”.<br>
We rephrased that paragraph, and the presentation of the previous<br>
measurements has been reordered.<br>
- "L 187-188: F1 and F2 already defined. Put a full stop at the end of Eq<br>
16, skip “where…factors” and write “Due to the relative values of the proton<br>
form factors F1 and F2,….”"<br>
Thanks, it is way better now.<br>
- "L190-191: this is not quite correct, and we had that discussion before<br>
for the PRL and checks were made with VGG. The contributions of H and<br>
H_tilde to A_UL are quite comparable. It is misleading to say that the<br>
latter is dominant and the former minor. Please find another formulation."<br>
Ok, we modified the text explaining better why there is a comparable<br>
sensitivity to ImH and ImHtilde.<br>
- "Section III: may be worth adding towards the end the beam helicity flip<br>
frequency and a typical target polarization reversal frequency."<br>
While we know that the helicity change is the usual one - i.e. 30 Hz - it is<br>
hard to define a typical frequency in changing the target polarization,<br>
since the changes were not regular at all. We preferred then not to mention<br>
this information, not to create "asymmetry" among the two polarization<br>
information.<br>
- "About combining results from parts A and B, it would be instructive (and<br>
maybe necessary) to add what is the expected beam energy dependence of the 3<br>
observables from a given model (say VGG), for fixed xB, Q2, t. It is not<br>
enough, as is done later, to show that the bin centers are nearly the same."<br>
We did some studies through the VGG model, the results of which can be found<br>
here:<br>
<a href="http://www.lnf.infn.it/~pisanos/dvcs/prd/coll_wide_review/ebeam_test/" target="_blank">http://www.lnf.infn.it/~<u></u>pisanos/dvcs/prd/coll_wide_<u></u>review/ebeam_test/</a><br>
The two curves are the -t dependence of the three asymmetries in our 5 (Q2,<br>
xb) bins for the two beam energies (blue line is partA, red partB). Since<br>
the difference is very small it is well beyond our statistical precision.<br>
- "Table III: same number of significant digits everywhere ?"<br>
To the best of our knowledge the significant figures are correct as is. For<br>
statistical errors one usually retains only the first non-zero digit unless<br>
the error is small (e.g. 0.015) and then it is common practice to keep two<br>
digits. For these reasons all the PbPt and one of the Pb errors have two<br>
digits, while the other ones have been rounded to one.<br>
- "L853: $R_Acc, 1.3\times R_Acc, 0.7\times R_Acc$"<br>
We explained it in words, to make the sentence more readable.<br>
- "L891 and others: we call GK (from the original GPD parameterization) a<br>
calculation of DVCS observables made by KMS (and for this paper by MS).<br>
Although this is a bit heavier notation, why not replacing GK by GK-KMS ? GK<br>
did not calculate DVCS, and it would be fair to our collaborators on this."<br>
Unfortunately, we cannot change it anymore since we used the same notation<br>
for the PRL, and we need to be consistent among the two papers.<br>
- "Fig.26 caption: is not it -ImH/pi and - ImH_tilde/pi which are plotted?<br>
If yes, should be explicit in the caption."<br>
The quantities that we plot in the figure are the actual imaginary parts of<br>
the CFFs, because we keep into account the -pi factors.<br>
- "Table V: strictly speaking, TSA, its errors and c (DSA and other c)<br>
should have the same number of significant digits (4). I would leave it as<br>
is, but make sure it is the case in the CLAS data base."<br>
We adopted the convention to present all the results with 3 sig figures,<br>
however being the TTSA correction so small we decided to report with 4 sig<br>
figures to avoid zero values.<br>
Thanks again!<br>
</blockquote>
<br>
--<br>
***<br>
Michel GUIDAL<br>
Institut de Physique Nucleaire<br>
Bat 100 - M019<br>
91406 ORSAY Cedex<br>
Tel: (33) 01 69 15 73 21<br>
Fax: (33) 01 69 15 64 70<br>
E-mail: <a href="mailto:guidal@ipno.in2p3.fr" target="_blank">guidal@ipno.in2p3.fr</a><br>
***<br>
</blockquote>
<br>
</blockquote>
<br>
-- <br>
***<br>
Michel GUIDAL<br>
Institut de Physique Nucleaire<br>
Bat 100 - M019<br>
91406 ORSAY Cedex<br>
Tel: (33) 01 69 15 73 21<br>
Fax: (33) 01 69 15 64 70<br>
E-mail: <a href="mailto:guidal@ipno.in2p3.fr" target="_blank">guidal@ipno.in2p3.fr</a><br>
***</div></div></blockquote></div><br><br clear="all"><div><br></div>-- <br><div class="gmail_signature"> "Life was just chemistry, <br> and chemistry was just physics." - Jonah Lehrer</div>
</div>