The things I think we should have ready for LCWS are:
o A detailed sensor simulation of the actual sensor submitted for
  fabrication to get the most realistic charge diffusion, etc, values.
  This needs to cover all the 21 points using in the physics simulation.
  Ideally, this would go out further that just the eight nearest
  neighbour pixels as there is some charge lost outside of the 3x3
  array. I assume Giulio is the only person who can do this.
o We need to show we understand the resolution for EM showers both for
  MAPS and for the diode pads, meaning where we hit the physics shower
  fluctuation and/or Si-W sampling fraction limits. I think this means
  determining the resolution for the ideal GEANT4 analogue case (where
  the resolution is then only due to the fluctuations and sampling
  fraction), and then building up effect-by-effect to see where the
  resolution we see comes from, until we get to the full MAPS digitisation
  and hit clustering.
o The resolution dependence on all the various assumptions of the model,
  e.g. noise, threshold, charge diffusion, dead fraction, etc. Again, this 
  must be done with full digitisation. We need to then decide on the
  nominal values for parameters such as noise and threshold to be used for
  other results.
o The linearity and resolution as a function of energy from 0.5GeV to
  500 GeV; this is most easily done for photons to get round any B field
  issues. This clearly needs to include all digitisation effects. 
  (Ideally, this would be done for various pixel sizes but I don't think 
  this will be trivial to do in the time before LCWS.)
o For PFA, then the most basic thing we want to show first is that the 
  MAPS option does no harm. I think the easiest way to proceed with this
  is to sum pixels over the area equivalent to the diode pads and take
  the energy to be proportional to the number of hit clusters within the
  diode equivalent area. By forming CalorimeterHits from the non-zero
  pads, then this can be fed directly into PandoraPFA and should give
  directly comparable results to the diode pad case with no further
  tuning or rewriting of the PFA algorithm. Any degradation (if any) in 
  the resolution of hadronic jets is then the worst case for MAPS and 
  sets the lower limit of the MAPS performance. Clearly, having done
  this, then varying the diode pad size (2x2mm2, 5x5mm2, 10x10mm2, etc)
  would be an interesting study. Obviously full digitisation is needed 
  for this to be meaningful.
o If we can then move to using the finer granularity of MAPS to improve
  on the diode pad equivalent PFA resolution, then that would be even 
  better but this should be a second priority to the above. (It is a 
  huge study and is unlikely to be fully completed by LCWS, so we should
  concentrate on things we can finish initially.)
o There are probably other items we should do so feel free to add to
  this list when we meet.