Alex Vlassov thinks he has found a problem with
the calorimeter implementation in GSIM. He believes
he has found instances where the attenuation lengths
are set to 10**-2. In looking at the new attenuation
length file, there are a number of attenuation lengths
which are 0.0, but none which are 10**-2. An attenuation
length of 0.01 may cause a floating point exception for
short distances. Presumably there is a trap for zero
attenuation lengths; if so there is either some unlikely
roundoff error or an error in reading in the attenuation
lengths. Maybe someone could look into this.
Alex also believes he has found instances of negative
distances from the fiber readout. This might be possible
if the distance is taken from the edge of the triangle since
there is a rectangular piece which extends beyond the triangle
shape. All the negative distances he has seen are small.
- Will