July 3, 1998

$MCFAST_DIR/example/simulator/advanced_example_1
$MCFAST_DIR/example/simulator/advanced_example_2
$MCFAST_DIR/example/simulator/advanced_example_3

In addition there is one unsupported mixing example: ~bphyslib/mixing_example

The last example is present on the fnalu cluster ( and, for developers only, on fnsimu1 ).

The detailed documentation for each example is found in README files which are in the indicated directory heirarchy.

• Advanced Example 1
  This example exercises things like vertex fitters, mass fitters, and the mcfast utilities, such as code to compute the error on a proper decay time, code to compute the error on an impact parameter etc. The example uses the decay:
  B_s-> psi K*bar -> mu+ mu- K- pi+.
  The code peeks at the generator information to find out which tracks should be assigned to the decay in question and which to the main vertex. It then starts the exercise of the various fitters and utilities. The code also contains many illustrative histograms which show that the utilities are actually working - ie some distribtions should be unit gaussians and other should be a flat probability of chi2. Finally the example source includes the histogram file which came from a good run.

• Advanced Example 2
  In this example the evgen job is told not to decay any hadrons containing heavy quarks. These are decayed within mcfast. The example contains some histograms which show that neutral heavy hadrons ( D0, B0, B_s) have not turned in the magnetic field but that the charged heavy hadrons have.

  In this example, the K0s and stable hyperons which are produced inside evgen still have zero lifetime.

• Advanced Example 3
  This example shows how to write a chain of evgen jobs in which the random number seed is propagated from one job to the next. At present this example only works with pythia, not with isajet or herwig.

  A minor and, for the time, ignoreable weakness of this example is that the random number seeds for qq and for mcfast are not propagated from one job to the next -- the are initialized in each job using the time of day.

• Unsupported Mixing Example
  Consider doing an analysis of B_s mixing in which all particles are allowed to mix. ( Unlike the common case in which mixing is turned off in the simulation and then restored "by hand" ). In a significant fraction of the events, the B_s being looked at will be accompanied by another B_s or B_sbar - since the B_s system is almost fully mixed, there will be many events in which two B_s mesons ( and no B_sbar ) are produced.

  Now consider writing the event generator to simulate the above process and to study a specific channel of the B_s. In this case, one wants to force the first B_s meson in an event to decay into the channel being studied. The second B_s meson should be allowed to decay according to the default branching ratios.

  Now suppose that the channel in question is a long decay chain, such as:
  B_s -> Psi K*bar -> mu+ mu- K- pi+.
  It is important to only force the psi and K* decays if the psi and K* are decay products of the first B_s in the event.

  This example does all of the above. It has the added bonus that the correct particle ID codes are preserved at all steps in the job. If you look at this example, you will see that it requires changes to both qq and stdhep. For this reason it will remain unsupported until those changes are incorporated into the released versions of qq and stdhep.