Shower merging

The calorimeter response simulation does not consider the problem of merging of showers close together. Since this is a rather important feature in eg. deciding if a neutral EM shower is a single gamma or two gammas from a $\pi^0$ decay, a routine (ZAUSHO) to be called in the analysis part of SGV is supplied to do this. The showers are merged if the separation of the two showers (separation meaning separation in the four shower-axis parameters simultaneously) is such that the probability is greater than some user-defined value that two independent measurements of the SAME shower would give that same observed separation (or less). (Hence, big limiting probability implies many showers merged, and v.v.). This is the multi-dimensional analog to the well known way to decide if two single-number values will be possible to separate when measurement precision is taken into account, ie. by checking the 'number of sigmas' separating the true values. Merged showers will have a seen energy equal to the sum of the those of the contributing showers, and position and direction equal to the weighted average of them. Note that merging is done only within a calorimeter, not between different ones. Hence, the possible m.i.p. signal from a hadron in an electro-magnetic calorimeter will not be merged with the shower from the same hadron in the hadron calorimeter behind.