============================================
Minutes of Tracking Meeting, August 22, 2007
============================================

taken by U. Husemann


Slava Krutelyov: COT t0 Calibration
====================================

- Goal: get t0 right within approx. 200 ps (for analyses using track timing)
- Study shown today: get t0 calibration for Gen7 data
- Method: 
  o Time-like residuals between good tracks and good hits 
    (see slides for details of selection code)
  o Take single long run as reference to calibrate others against
- Samples: compare 6.1.4int10 (Z->ee, Jet20) vs. default production
  o Miscalibration in cell 0 of a superlayer is gone in int10
  o Similar spread of <t0> for all other cells (around 1 ns)
  o Shifts in <t0> are approximately sample-independent
  o Shifts in <t0> are slightly run-dependent
  o Misalignment in some cells, seen both in time-like and space-like residuals
- Alignment:
  o Slava's study and all validation so far: Gen 6 alignment
  o New COT alignment expected within 1 week (cosmics alignment available,
    needs additional beamline corrections)
  o Final t0 numbers need Gen7 alignment
- Cell-by-cell vs. wire-by-wire calibration:
  o Slava's code produces both calibrations: can be cross-checked
  o Uncertainty within cell smaller than between cells
  o Will check for common patterns across cells, which could be		
    incorporated in correction


All: Discussion on Gen7 Validation
==================================

- Alignment:
  o COT alignment expected next week, Silicon alignment to follow
  o Ideally: new alignment files go into Gen7 pre-release
  o Final validation effort will use new alignment
- Questions & concerns from Ulrich's Joint Physics talk:
  o Are additional hits on tracks of good quality?
  o Execution speed
- More on Execution speed:
  o Antonio's study: 66 seconds for 200e30 min bias data, 
    4 seconds for low luminosity data
  o How does old code compare to new code at 200e30?
  o What are the expected limitations for execution speed (offline group)?
  o Need reduction of combinatorics in Silicon standalone tracking:
    pT cut on the pattern recognition level (to be studied)
- Antonio is processing high-pt lepton data from one 0i period with int10:
  o Tom interested in production output
  o (Wisconsin group interested, too)