class AliTRDresolution: public AliTRDrecoTask

 Default constructor

 Default constructor

 Destructor

 spatial resolution

 Init containers for subsequent tasks (AliTRDclusterResolution)

 Execution part

 Helper function to calculate pulls in the yz plane
 using proper tilt rotation
 Uses functionality defined by AliTRDseedV1.

 Plot the cluster distributions

 Plot the cluster distributions

 Plot normalized residuals for tracklets to track.

 We start from the result that if X=N(|m|, |Cov|)


 in our case X=(y_trklt - y_trk z_trklt - z_trk) and |Cov| = |Cov_trklt| + |Cov_trk| at the radial
 reference position.

 Store resolution/pulls of Kalman before updating with the TRD information
 at the radial position of the first tracklet. The following points are used
 for comparison
  - the (y,z,snp) of the first TRD tracklet
  - the (y, z, snp, tgl, pt) of the MC track reference

 Additionally the momentum resolution/pulls are calculated for usage in the
 PID calculation.

 Plot MC distributions

 Get range on Z axis such to avoid outliers

 Get the reference figures

 Build summary plots

 Draw error bars scaled with "scale" instead of content values
use range [m,M] if limits are specified

 Returns the range of the bulk of data in histogram h2. Removes outliers.
 The "range" vector should be initialized with 2 elements
 Option "mod" can be any of
   - 0 : gaussian like distribution
   - 1 : tailed distribution

 Analyse cluster

 Analyse cluster

 Analyse tracklet

 Analyse track in

 Analyse tracklet

 Fit, Project, Combine, Extract values from the containers filled during execution

 Helper function to avoid duplication of code
 Make first guesses on the fit parameters

 Define histograms

 Fit track with a staight line using the "np" clusters stored in the array "points".
 The following particularities are stored in the clusters from points:
   1. pad tilt as cluster charge
   2. pad row cross or vertex constrain as fake cluster with cluster type 1
 The parameters of the straight line fit are stored in the array "param" in the following order :
     param[0] - x0 reference radial position
     param[1] - y0 reference r-phi position @ x0
     param[2] - z0 reference z position @ x0
     param[3] - slope dy/dx
     param[4] - slope dz/dx

 Attention :
 Function should be used to refit tracks for B=0T

 Fit tracklet with a staight line using the coresponding subset of clusters out of the total "np" clusters stored in the array "points".
 See function FitTrack for the data stored in the "clusters" array

 Global selection mechanism of tracksbased on cluster to fit residuals
 The parameters are the same as used ni function FitTrack().

 Get the most probable value and the full width half mean
 of a Landau distribution

 steer processes

 steer processes

  Bool_t          HasTrackRefit() const                 { return TestBit(kTrackRefit);}
Bool_t          HasTrackSelection() const             { return TestBit(kTrackSelect);}

  void            SetTrackRefit(Bool_t v = kTRUE)       { SetBit(kTrackRefit, v);}
void            SetTrackSelection(Bool_t v = kTRUE)   { SetBit(kTrackSelect, v);}