new release of REZEST

["Mark M. Ito" <marki@jlab.org>]

I've made some changes to the resolution estimation code, prompted by 
discussion at last Friday afternoon's tracking resolution meeting:

1. Introduced a new parameter specifying the minimum number of hits 
needed in a detector for it to provide a measurement. Currently set to 3.
2. Changed the method for calculating the curvature error contribution 
to the azimuthal angular error. I was using a fixed radius for the 
detector position; now the position depends on the direction of the track.
3. Improved the calculation of the curvature error contribution mention 
in (2). Use an exact formula rather than a small angle approximation.
4. Introduce material in front of the FDC equal to that in front of the 
CDC. In addition add the material in the downstream endplate of the FDC.

The agreement is quite a bit better now in both azimuthal and polar 
angular resolutions at small angles.

The svn revision is 3498 and the tar file and instructions mentioned on 
the wiki page at

http://www.jlab.org/Hall-D/software/wiki/index.php/Tracking_resolution_estimator_%28REZEST%29

have been updated.