Minutes for Simulation Working Meeting
          ======================================

Meeting time:  Sept. 26, 2001 1:00-3:00pm
Meeting place:  9th Circle
Attend:  Laurent Duflot, Marek Zielinski , Elizaveta Shabalina, 
         Serban Protopopescu, Qizhong Li, Don Coppage, Dookee Cho, 
         Vishnu Zutshi, Anna Goussiou

Agenda:
        This is a work meeting to understand the
        plate vs mixture in p10 production release.


This working meeting is very fruitful in understanding the problems
with mixture vs plate in p10.

	- Serban showed his plots of PT distribution from single electron.
          The one with mixture looks fine. But the one with plate geometry
          gives <PT> of 200 something for 20 GeV electron. 
          
          This turns out to be not a problem in p10. Serban generated 
          these single electron events using p09.08.00 d0gstar, t156 d0sim,
          t161 emreco and em_analyze. He didn't use p10 d0sim to generate 
          these single events.  He used t156 d0sim. The cal weight has
          been applied twice in t156. This is fixed in p10.
          Laurent explained why applying cal weight twice will get
          Serban's results. The mixure has weight of 1. So applying the
	  weight twice will not change the <PT>. But the plate weight is 
          about factor of 3 more. So applying twice plate weight will
          make the <PT> factor of 10 larger.
	  Vishnu took Serban's singl electron d0gstar output, run through
          p10 d0sim/d0reco. The events look OK. The <PT> plot has no
          problem in plate MC. <PT>=17.24 GeV for 20 GeV electron.

        - Understand the differences for total energy in EM1 between
          plate and mixture. From Armand's plots we see that the
          distribution for pT in a cell looks very similar for mixture 
          and plate, but the # of hits are 15% more in EM1 for plate
          than for mixture and the total E in cal is 10 GeV larger
          in EM1 for plate than for mixture.
          We do expect differences between mixture and plate. The plate
          geometry is more accurately simulating the D0 geometry. 
          But we want to make sure we didn't have mixtake or bug in
          plate to cause the difference.
          Laurent checked all the cal weights used in p10. They are
          confirmed to be OK.
          In EM1, we do expect bigger difference because mixture has 
          two uranium layers and plate has one uranium layer with two
          argon layer.
          The number of hits in a layer should be different because
          in p10 the ECUT is different for mixture and plate. The
          mixture uses 1 MeV cut for gamma and the plate uses 10 KeV
          cut for gamma.

	- Laurent pointed out that p09 uses very old (1998?) plate 
          weight and p10 uses the newly determined cal-only weight for
          plate.
          Current EMRECO gives wrong weight for electron in real data, 
          because no cps in real data.
          In general, the jets prefer cal-only weight and the electrons
          prefer cal+cps weight.


We also discussed the p11 production release.
        
        - The dead line for putting new feature into p11 is Oct. 15.
          After that date, only bug fixes are allowed.
          cpb required d0gstar/d0sim/d0trigsim to get ready two weeks
          before d0reco.

	- We should start to check D0Raw2Sim. The program is already
          released in p10. But we need to do intergrated test, especially
          we need to test it on the real data.

	- Anna offered to pin two zero-bias runs raw data (taken yesterday)
          on disk, so we can test D0Raw2Sim using the zero-bias real data.