[Halld-offline] [Halld-physics] Next Data Challenge...

[Matthew Shepherd]

Hi Dave,

To add to this:

On Jul 1, 2012, at 11:30 PM, David Lawrence wrote:

> Our plan up to now, as I understand it, is to produce and analyze simulated data and store only the DST (or REST format as described at the last software meeting). I think that should be done as a separate exercise from the real data Data Challenge exercise.


If EVIO format is not available, I would suggest we produce and store a large data set of either hdgeant or mcsmear output format (we can discuss which).  One step of the data challenge will be making the REST format data from this sample.  This involves running and managing many instances of some standard JANA-based job to generate the REST files.  This will be analogous to what is done when the raw data are reconstructed except the input will be a different format.  This should definitely be done at JLab on the JLab farm since it is exactly how we plan to do the reconstruction.

The second step should be to conduct two or more simultaneous analyses using the REST output without access to the raw hits or having to run the reconstruction code.  Again, the skim should probably be done at JLab rather than shipping all REST format offsite.  One may consider making a loose skim at JLab and then moving those files offsite for further analysis.

For now, I think developing an algorithm with kinematic fitting, etc., that runs in the jana framework to isolate some exclusive channel in the PYTHIA sample is sufficient to count for an "analysis."  We know it is the signal/background ratio that we are most concerned with and that is what can be readily tested using these large PYTHIA samples.  In other words, I don't think for this challenge we need to blend in signal MC with amplitudes and conduct an amplitude analysis.

Matt