Tracker Recon Software Meeting
July 14, 2004
(Present: Bill, Leon, Tracy)
- Towards a new Fit Track Class
- A start of an overview of the structure and contents of a new fit track
TDS class is located here
- Some of the resulting discussion:
- Output of pattern recognition will be the fit track class - no pattern
rec only class
- Criteria for keeping a variable in the output class:
- How many places in the code is it used (eg display, vertexing,
analysisntuple...)
- Alternative is to create helper classes which calculate important
but infrequently used quantities
- Class should be organized to provide forward and backward iterators
over the plane hits
- Want to remove explicit loops over plane numbers in code to allow
for upward/downward track fits
- Vertexing/Track Topology
- Need to begin to look at what has been called vertexing in TkrRecon
- Will probably abandon Kalman Filter vertexing? Not applicable to our
problem?
- Bill would like to go to event topology versus "vertexing" in the
HEP tracking sense
- A plan for the rest of the summer's meetings:
- No meeting July 22, both Bill and Tracy will be out of town
- July 29: Finalize the TDS classes for clusters and fit tracks
- August 5: Make existing code work with these new classes
- August 12&19: Tracy unavailable for the first, Bill for both
- August 26: First go round on a revamped TkrRecon Gaudi architecture
- September: Begin the update of the TkrRecon code - Goal: 9/19 push the
button
- Action items
- Remaining from previous meeting:
- Leon and Bill to make a list of how clusters are accessed in current code,
and what other access modes would be desirable.
- Leon to continue looking at how to improve TkrCluster and access.
- From this meeting:
- Bill will send us a list of critical items to be done before the next
round of work on background rejection
- Bill will start thinking about "vertexing" or an event topology stage
after track fitting
- Tracy to complete the new fit track class overview
T.Usher Last Modified:
07/14/2004 19:44