class AliFMDReconstructor: public AliReconstructor

 Make a new FMD reconstructor object - default CTOR.

 Destructor

 Initialize the reconstructor

 Convert Raw digits to AliFMDDigit's in a tree

 Return the vertex to use.
 This is obtained from the ESD object.
 If not found, a warning is issued.

 Get the detector identifier.
 Note the actual value is cached so that we do not
 need to do many expensive string comparisons.

 Get the reconstruction parameters.

 Return:
   Pointer to reconstruction parameters or null if not found or wrong type

 Set-up reconstructor to use values from reconstruction
 parameters, if present, for this event.   If the argument @a set
 is @c false, then restore preset values.

 Parameters:
    set

 Loop over all entries of the ESD and mark
 those that are dead as such
 - otherwise put in the zero signal.

 Reconstruct directly from raw data (no intermediate output on
 digit tree or rec point tree).

 Parameters:
   rawReader	FMD Raw event reader

 Reconstruct directly from raw data (no intermediate output on
 digit tree or rec point tree).

 Parameters:
   reader	Raw event reader
   ctree    Not used - 'cluster tree' to store rec-points in.

 Reconstruct directly from raw data (no intermediate output on
 digit tree or rec point tree).

 Parameters:
   reader	Raw event reader
   ctree    Not used.

 Reconstruct event from digits in tree
 Get the FMD branch holding the digits.
 FIXME: The vertex may not be known yet, so we may have to move
 some of this to FillESD.

 Parameters:
   digitsTree	Pointer to a tree containing digits
   clusterTree	Pointer to output tree

 For each digit, find the pseudo rapdity, azimuthal angle, and
 number of corrected ADC counts, and pass it on to the algorithms
 used.

 Parameters:
    digits	Array of digits

 For each digit, find the pseudo rapdity, azimuthal angle, and
 number of corrected ADC counts, and pass it on to the algorithms
 used.

 Parameters:
    digits	Array of digits

 Process a single digit

 Parameters:
    digit Digiti to process

 Process the signal from a single strip

 Parameters:
    det	Detector ID
    rng	Ring ID
    sec	Sector ID
    rng	Strip ID
    adc     ADC counts

 Process the signal from a single strip

 Parameters:
    det	Detector ID
    rng	Ring ID
    sec	Sector ID
    rng	Strip ID
    adc     ADC counts

 Subtract the pedestal off the ADC counts.

 Parameters:
    det           Detector number
    rng           Ring identifier
    sec           Sector number
    str           Strip number
    adc           ADC counts
    noiseFactor   If pedestal substracted pedestal is less then
        this times the noise, then consider this to be 0.
    zsEnabled     Whether zero-suppression is on.
    zsNoiseFactor Noise factor used in on-line pedestal
        subtraction.

 Return:
    The pedestal subtracted ADC counts (possibly 0), or @c
         USHRT_MAX in case of problems.

 Member function to subtract the pedestal from a digit

 Parameters:
    det	Detector ID
    rng	Ring ID
    sec	Sector ID
    rng	Strip ID
    adc     # of ADC counts
 Return:
    Pedestal subtracted signal or USHRT_MAX in case of problems

 Converts number of ADC counts to energy deposited.
 Note, that this member function can be overloaded by derived
 classes to do strip-specific look-ups in databases or the like,
 to find the proper gain for a strip.

 In the first simple version, we calculate the energy deposited as

    EnergyDeposited = cos(theta) * gain * count

 where

           Pre_amp_MIP_Range
    gain = ----------------- * Energy_deposited_per_MIP
           ADC_channel_size

 is constant and the same for all strips.

 For the production we use the conversion measured in the NBI lab.
 The total conversion is then:

    gain = ADC / DAC

                  EdepMip * count
      => energy = ----------------
                  gain * DACPerADC

 Parameters:
    det	Detector ID
    rng	Ring ID
    sec	Sector ID
    rng	Strip ID
    counts  Number of ADC counts over pedestal
 Return
    The energy deposited in a single strip, or -1 in case of problems

 Converts number of ADC counts to energy deposited.
 Note, that this member function can be overloaded by derived
 classes to do strip-specific look-ups in databases or the like,
 to find the proper gain for a strip.

 In the first simple version, we calculate the energy deposited as

    EnergyDeposited = cos(theta) * gain * count

 where

           Pre_amp_MIP_Range
    gain = ----------------- * Energy_deposited_per_MIP
           ADC_channel_size

 is constant and the same for all strips.

 For the production we use the conversion measured in the NBI lab.
 The total conversion is then:

    gain = ADC / DAC

                  EdepMip * count
      => energy = ----------------
                  gain * DACPerADC

 Parameters:
    det	Detector ID
    rng	Ring ID
    sec	Sector ID
    rng	Strip ID
    eta     Psuedo-rapidity
    counts  Number of ADC counts over pedestal
 Return
    The energy deposited in a single strip, or -1 in case of problems

 Converts an energy signal to number of particles.
 Note, that this member function can be overloaded by derived
 classes to do strip-specific look-ups in databases or the like,
 to find the proper gain for a strip.

 In this simple version, we calculate the multiplicity as

   multiplicity = Energy_deposited / Energy_deposited_per_MIP

 where

   Energy_deposited_per_MIP = 1.664 * SI_density * SI_thickness

 is constant and the same for all strips

 Parameters:
    det	Detector ID
    rng	Ring ID
    sec	Sector ID
    rng	Strip ID
    edep    Energy deposited in a single strip
 Return
    The "bare" multiplicity corresponding to the energy deposited

 Get the eta and phi of a digit

 Parameters:
    det	Detector ID
    rng	Ring ID
    sec	Sector ID
    rng	Strip ID
    eta	On return, contains the psuedo-rapidity of the strip
    phi     On return, contains the azimuthal angle of the strip

 nothing to be done
 FIXME: The vertex may not be known when Reconstruct is executed,
 so we may have to move some of that member function here.

 Forwards to above member function

   * CTOR

   * Flag that we can convert raw data into digits.
   *
   * @return always @c true

   * Return the filled FMD ESD object
   *
   * @return FMD ESD object

   * Not used

   * Set the noise factor
   *
   * @param f Factor to use

   * Set whether we should do angle correction or nor
   *
   * @param use If true, do angle correction

   * Set whether we want to do diagnostics.  If this is enabled, a
   * file named @c FMD.Diag.root will be made.  It contains a set of
   * histograms for each event, filed in separate directories in the
   * file.  The histograms are
   * @verbatim
   * diagStep1   Read ADC vs. Noise surpressed ADC
   * diagStep2   Noise surpressed ADC vs. calculated Energy dep.
   * diagStep3   Energy deposition vs. angle corrected Energy dep.
   * diagStep4   Energy deposition vs. calculated multiplicity
   * diagAll     Read ADC vs. calculated multiplicity
   * @endverbatim
   *
   * @param use If true, make the diagnostics file

   * Assignment operator
   *
   * @param other Object to assign from
   *
   * @return reference to this object