[Halld-offline] Modified DDetectorMatches proposal

[Paul Mattione]

After some discussions I've decided to modify the proposal; the new header file is attached.  Now for each match between a track and a BCAL/FCAL/TOF/SC shower/hit, it would save the following info to REST:

- dx OR dE/dx (for shower/hit) (dx is distance along the track through the BCAL/FCAL to the shower position)
- flight time to hit/shower
- flight time variance
- path length to hit/shower
- track DOCA to hit/shower

This should be enough information for us to determine during the analysis stage which shower/hit is the "best" match we should select to calculate the PID FOM, etc. without needing to re-swim the tracks.  Or perhaps we could develop a future algorithm to do something more complicated with it. 

To build the full covariance matrix for kinematic fitting with timing constraints, it will also save the correlation coefficients between the track momentum and the flight time to each system.  However, it currently assumes that this correlation is unchanged between different objects of the same type (e.g. between different BCAL showers).  This is because I think calculating this correlation to each object will be computationally expensive, although I'm still working with Simon on that.  

Finally, it will also separately save the distance from each shower to the closest track (regardless of matching), as Will showed earlier that it might helpful for separating neutrals from hadronic showers in the BDT.  

Please let me know if you guys have any more comments/suggestions for changes.  

 - Paul

-------------- next part --------------
A non-text attachment was scrubbed...
Name: DDetectorMatches.h
Type: application/octet-stream
Size: 3015 bytes
Desc: not available
URL: <https://mailman.jlab.org/pipermail/halld-offline/attachments/20131218/fa9cd065/attachment-0002.obj>