Minutes of the muon software meeting Wednesday 3rd September 2003, taken by Victoria. Talks are on the web at: http://www-cdf.fnal.gov/internal/WebTalks/0309/030903_muon_software.html 0. Announcements ----------------- David has a new job as production coordinator and Lucio Cerrito is replacing him co-chair of the muon software group. We all gave David a round of applause to thank him for his work, and for continuing to find bugs in the muon software. 1. Victoria: Updates to the software. -------------------------------------- Two changes have been made to the software in the last few weeks. The first is the addition of inside-out and SVXStandAlone tracks in the muon, and other was a bug fix effecting the MuonMatchData associated to some muons. A few plots were shown showing the differences between the (old) 4.11.1 processing and a 5.1.0pre6 + muon fixes processing. The most worrying change seems to be in the stub type attached to the tracks. There are more stubless muons and less CMUP and CMX muons. (Post-meeting I realized that I didn't make a cut on d0. If I do make that cut the distribution looks like we would hope: same number of CMUP muons, slightly more CMX muon, many more stubless guys). It has also been noticed that there are many more stubless muons with zero associated energy. This seems mainly due to the new track types used in muons. Comments: many people suggested to look at the matching variables to see what the effect of the MuonMatchData bug was. 2. Slava: update on CMX chi2. ------------------------------ Slava showed an update of his work to formulate a chisq variable for CMX muons using J/psi->mu+mu- events. Using his old formula for the mean of dxCMX he has obtained a function for the width of dxCMX. This function is second order in pT and eta. Using a cut of chisq=(dxCMX/sigma_dx)^2 < 9 the efficiency for pT>3 is 0.9977 +/- 0.0001. Comments: It was not clear in the efficiency as a function of pT was smooth: the 2-d colour plots were hard to interpret. Lucio also asked about the dxCMX as a function of eta. The function Slava formulated still shows an eta dependence even at high pT. In his work Lucio had seen that there was no eta dependence for pT>8 GeV/c. 3. Round table discussion -------------------------- We had a discussion about work that still needs to be done on the muon software. Ken said the current version of the cosmic tagger tags too many events as cosmics. It is also biased towards tagging high-multiplicity events as cosmics. There has been some work done on this on the W and Z cross section group, and this could be discussed in the context of updating the tagger code. Liz confirmed that there would be a physics code release (after the production release) and so any adjustments to the cosmic tagger could go there. We also discussed the work that needs to be done on the simulation and the geometry. Michael Gold (with John Strologas) will work on improvements to the simulation. Lucio described what work needed to be done to improve the simulation: the matching variables are very different between data and MC, especially at low pT. David also commented that the BMU containment volume overlaps the miniskirt in the geometry. This will need to be fixed before the miniskirt and/or the BMU simulation works properly. Victoria read out the rest of her to do list, which includes updating the validation modules in order to make a good run list. Writing classes to tag the trigger information and to perform a set of standard muon cuts. Plus Phoenix tracking for BMU muons (which Dan is working on). David commented that the numbers he got for the BMU reconstruction efficiency didn't make any sense between data and MC, and therefore he thinks there is still work to do in the BMU reconstruction. David also announced that there is going to be a change in format of the raw TDC in the post (October) shutdown data. New data won't be reconstructible with old code. This shouldn't effect any of the muon code. The next meeting is scheduled for Wednesday 17th. However as this is a day before the collaboration meeting we may cancel it.