Here is a list of things to be done, volunteers (as usual) are welcome

• Sep 19 1998
  • Simulated primary vertices: need to write this information out
  • MC_EVENT structure: MC_PARTICLE destructor should not "know" about the detectors, as of today - it does. Need to restructure the code.
• Detailed calorimetry simulation: John Mayer has a standalone GEANT simulation for plug. Has to be integrated. No detailed simulation for the central calorimeter is available.
• fast tracking simulation: do we need simplified description of the tracker geometry at all or the level of details we have is quite enough?
• Optimization of memory management during hit generation
• Secondary vertices: decay charged strande baryons (lambda's, sigma's) etc taking into account magnetic field
• neutrino's : do we want to pass them to GEANT ?
• writing out SVX MC banks with MC truth information in there (equivalent of MCOT)
• Control structures: how to handle switches (where to keep them, how to name etc)
• Interactive input: what do we want

- as a short term solution
- in a long run
• A particle in GEANT may interact with the detector material several times (i.e. the same electron emits several brem photons) In this case particle has several "decay" vertices, this destroys the correspondence between particles and vertices in MC_EVENT ... This doesn't hurt tracing however, but some of the vertices have parent which doesn't stop in this vertex. Should be fixed somehow.. The simpliest way: terminate particle when it brems and create 2 daughters: scattered particle and brem photon Another option: particle may have a list of interaction vertices ??? What about the decay vertex ?
• Random numbers: code should be debuggable. Suppose we generated 1000 events and know starting random numbers for the event #1000. If we initialize random number generator(S ?) with these numbers, the 1st simulated event should coinside with the event #1000 from the previous run. It is not the case right now.
• Random numbers again: we have to use separate stream for smearing the primary vertices - primary vertex position for event#10 shouldn't depend on the simulated geometry

DONE: