About CERN-PH-EP-2010-009:
Apologies for being long-winded, but I think CMS pubs are getting into a bad habit. At least for future publications we should be more careful.
Maybe(?) this paper is too far gone for this comment to have much effect, except perhaps for future analyses, but I have a real objection to our attempt to classify (inelastic) interactions into distinct classes: non-single-diffractive and single diffractive or double diffractive Which, by the way, is totally different (orthogonal) to double pomeron exchange. Double diffractive usually means X-G-X where X = a multiparticle state and G means a big ( >~ 4 or so, nothing specific, which is the key point) gap, while double pomeron exchange is pGXGp (in order of rapidity of course). Note also that pomeron should have a small p, it is not a proper name like Fermi or Bose, but a derivative like fermion and boson. I know it is often written with P but that is incorrect. (Higgs is not a particle, Higgs boson is OK).
Diffractive and non-diffractive interactions are treated differently theoretically but as experimenters there is in principle no absolute way to distinguish them, even with a 4π detector, let alone with CMS which has no forward detection (ZDC is not helpful for this). With a truly 4π detector we could say either (1) we choose to call an event single diffractive if there is a proton with Feynman xF > 0.95 (say), but why not 0.96 or 0.97, the transition is continuous, indeed in Regge theory there is interference between pomeron and reggeon exchange. However in CMS we do not even detect the proton so we cannot even apply this, purely experimental, criterion. Or (2) we could choose to call an event single diffractive if there is a leading rapidity gap, with no hadrons, over 3 units, but why not 2.5 or 3.5? Again there is no real distinction, one could choose any value in this region and make it an experimental definition. However in CMS we have no detector coverage between |η| = 5.2 (HF) and the beam rapidity (y = 8.2 at sqrt{s}=7 TeV) so we have no idea if there are particles there or not.
So we have absolutely no experimental way of separating events into such distinct classes. We should not. We can say we select inelastic events that have particles in both HF regions (which still accepts events with a gap of 3-4 units and might be said to be diffractive), and then include all such events in our central distributions. But it seems in this paper we also required hits in both BSC hodoscopes, which if true only picks out some fraction (what fraction?) of those. It is fair enough if we state clearly how we chose some subset of the inelastic collisions, but any attempt to define diffractive or non-diffractive subsets can only be dependent on what some theoretical or phenomenological models (e.g. PHOJET or PYTHIA say, and none of them do it properly (amplitudes!). It is time MC-builders took a holistic view of all inelastic collisions (some try) and we should encourage that.
Apologies for being long-winded, but I think CMS pubs are getting into a bad habit. At least for future publications we should be more careful.
Maybe(?) this paper is too far gone for this comment to have much effect, except perhaps for future analyses, but I have a real objection to our attempt to classify (inelastic) interactions into distinct classes: non-single-diffractive and single diffractive or double diffractive Which, by the way, is totally different (orthogonal) to double pomeron exchange. Double diffractive usually means X-G-X where X = a multiparticle state and G means a big ( >~ 4 or so, nothing specific, which is the key point) gap, while double pomeron exchange is pGXGp (in order of rapidity of course). Note also that pomeron should have a small p, it is not a proper name like Fermi or Bose, but a derivative like fermion and boson. I know it is often written with P but that is incorrect. (Higgs is not a particle, Higgs boson is OK).
Diffractive and non-diffractive interactions are treated differently theoretically but as experimenters there is in principle no absolute way to distinguish them, even with a 4π detector, let alone with CMS which has no forward detection (ZDC is not helpful for this). With a truly 4π detector we could say either (1) we choose to call an event single diffractive if there is a proton with Feynman xF > 0.95 (say), but why not 0.96 or 0.97, the transition is continuous, indeed in Regge theory there is interference between pomeron and reggeon exchange. However in CMS we do not even detect the proton so we cannot even apply this, purely experimental, criterion. Or (2) we could choose to call an event single diffractive if there is a leading rapidity gap, with no hadrons, over 3 units, but why not 2.5 or 3.5? Again there is no real distinction, one could choose any value in this region and make it an experimental definition. However in CMS we have no detector coverage between |η| = 5.2 (HF) and the beam rapidity (y = 8.2 at sqrt{s}=7 TeV) so we have no idea if there are particles there or not.
So we have absolutely no experimental way of separating events into such distinct classes. We should not. We can say we select inelastic events that have particles in both HF regions (which still accepts events with a gap of 3-4 units and might be said to be diffractive), and then include all such events in our central distributions. But it seems in this paper we also required hits in both BSC hodoscopes, which if true only picks out some fraction (what fraction?) of those. It is fair enough if we state clearly how we chose some subset of the inelastic collisions, but any attempt to define diffractive or non-diffractive subsets can only be dependent on what some theoretical or phenomenological models (e.g. PHOJET or PYTHIA say, and none of them do it properly (amplitudes!). It is time MC-builders took a holistic view of all inelastic collisions (some try) and we should encourage that.
Mike Albrow