<div dir="auto">Andrii, when in a hole it's a good idea to stop digging. Marian did not reject your help, he tried it and found (for his configuration, which can maybe be refined) that it did not work. In any case, the problem was in an aggressive and personalised response -- which is *never* appropriate -- rather than in the scientific merits which I'm sure we can debate at length without issue. I suggest we move on, with the rules of conduct now made explicit.<div dir="auto"><br></div><div dir="auto">On the question at hand, obviously the methodology was not designed with fiducial reinterpretation in mind, so we have to do something approximate. I have coded up a "generator definition" using fs leptons directly connected to the hard scattering -- which can be done safely. I hope this is unambiguous in the vast majority of cases. In case of ambiguity, it chooses the highest energy lepton. Would you be able to test this with your setup, Andrii? Yes it's more work (but that's what we do, isn't it?) but I for one would be very grateful, as we have been before for your contributions. And I don't think this is a circular and endless argument; actually, it feels to be like we have something closer to workable.</div><div dir="auto"><br></div><div dir="auto">Andy</div><div dir="auto"><br></div><br><div class="gmail_extra" dir="auto"><br><div class="gmail_quote">On 17 Nov 2017 07:44, "Andrii Verbytskyi" <<a href="mailto:andrii.verbytskyi@desy.de">andrii.verbytskyi@desy.de</a>> wrote:<br type="attribution"><blockquote class="quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Dear Leif,<br>
<div class="quoted-text"><br>
On Fri, 2017-11-17 at 11:10 +0100, Leif Lönnblad wrote:<br>
> On 2017-11-16 12:26, David Grellscheid wrote:<br>
> > please accept my apologies on behalf of the Rivet authors for this<br>
> > email. This is not an appropriate response to your question<br>
><br>
> I certainly agree with this! The tone in that letter was completely out<br>
> of line.<br>
<br>
</div>It is sad you think so. I've offered help to Marion to solve the problem<br>
but it was *explicitly* rejected. It does not look productive to me to<br>
invest my time in collaboration in this case. I will get more work and<br>
nobody will even say thanks.<br>
<div class="quoted-text"><br>
<br>
<br>
> On 2017-11-16 16:23, Andy Buckley wrote:<br>
> > My reading of the Rivet 2.5.3 code (copied below) is that it selects<br>
> > the highest-|pz| final-state lepton. Looking back through my emails I<br>
> > see that Andrii's concern about a final-state definition was actually<br>
> > about using highest-pT, i.e. not the same thing. DIS experts, is<br>
> > max-|pz| a reasonable heuristic?<br>
><br>
> Reading old H1 papers it seems their definition of the DIS kinematics is<br>
> based on their BEMC detector, which covers polar angles between 151 and<br>
> 177 degrees (wrt. the proton direction), and the scattered electron is<br>
> taken "as the most energetic BEMC cluster" with energy larger than 14<br>
> GeV. A rather simple experimental definition.<br>
<br>
</div>That is measurement technique and works well for low Q2. Not the case<br>
for high Q2 or charged current. Or even simpler -- what about<br>
central/back region?<br>
<div class="quoted-text"><br>
<br>
> However, looking at the old HZtool routines, they always use the<br>
> generator definition.<br>
<br>
</div>Exactly. And data was unfolded to it.<br>
<div class="quoted-text"><br>
> It's not clear to me if they generally correct for<br>
> this in any way. I saw one statement where they have studied QED<br>
> corrections and get effects of order 5% which they then include in the<br>
> systematic uncertainties, but do not correct for.<br>
<br>
</div>QED corrections is completely other thing. These are not included in<br>
Rivet anyway. This means generators should be used w/o FSR.<br>
<div class="quoted-text"><br>
> My guess is that any reasonable definition would do, and that the<br>
> original maximum-pz was reasonable, as is the generator definition. I<br>
> would, of course, prefer the the experimental definition, but it should<br>
> be easy to see what the differences are.<br>
<br>
</div>"Experimental definition" was quite complicated, i.e. in ZEUS it<br>
involved probabilities derived with neural network, isolation of<br>
electron, tracking and calorimeter information.<br>
<br>
<br>
Also, the whole discussion is going in circles (already 5 or more<br>
years). It is up to developers to decide and understand what they want,<br>
not up to me.<br>
<br>
<br>
Best regards,<br>
Andrii<br>
<br>
> /Leif<br>
><br>
><br>
<br>
<br>
</blockquote></div><br></div></div>