[Halld-offline] [GlueX] proposed contents of reconstructed MC dst

[Richard Jones]

This note summarizes the changes in gluex doc 1799 since it was first posted.  To leading order, it adopts all of the changes proposed by Paul Mattione and David Lawrence, with some modifications.  Read on for the details.


Draft 2 changes to the MC dst format since draft 1:

 1. *reaction (MC info header)* - removed "type", was included in case a Monte Carlo generator wanted to distinguish different types of reactions being merged together in the event stream.  We do not use this, and I don't foresee any demand for it.
 2. *vertex* - made clear that this is a list, whose average length for the benchmark MC sample used to meter the record size happens to be 1.  Added a unique "id" element for each vertex in the list, which is the referent of "vertexID" in the product section.
 3. *product *- took out the "pdgtype" field because it is redundant with the PID field, now renamed to "ptype".  Other field have been renamed, to better respect the rule that the keyword "id" indicate a relational handle.  This invented rule only applies to this document.
 4. *tagger hits* - made it explicit that this is a list of hits, and not just one, although current simulations feature only one per event.  Added a field for "id" in anticipation of kinematic fits objects we hope to add shortly that refer to a tagger hit.  Added a field for "time" (thanks David) but not for uncertainties.  Tagger energy uncertainties are too small to care about (3-5 MeV rms) and the time resolution will be a calibration db constant (or array) not a per-event quantity.  Right now it is unsmeared (zero error) in the simulation.
 5. *neutrals* - renamed this section DNeutralShowerCandidate (thanks Paul) to indicate the reconstruction object that it recapitulates.  The one difference between ours and the proposal for this section by Paul Mattione is the length of the "correlations" array.  We think there should be 3 off-diagonal matrix elements for Bcal showers and 2 for the Fcal.  Space is left for 3.  I will follow up this message with a new email detailing our arguments for this.
 6. *charged tracks* - renamed this section DChargedTrackHypothesis (thanks Paul) to indicate the reconstruction object it recapitulates.  Here I pretty much carried over everything from Paul's proposal, which looks cleaner than our original one, and complete.  Thanks to David for clarifying exactly what the "projected time" meant.  We changed some of the names to ones that better represent the meaning of the variables, we think.

-Richard Jones
(for Igor S. and rtj)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mailman.jlab.org/pipermail/halld-offline/attachments/20110818/b4f6e5fa/attachment-0001.html 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 2757 bytes
Desc: S/MIME Cryptographic Signature
Url : https://mailman.jlab.org/pipermail/halld-offline/attachments/20110818/b4f6e5fa/attachment-0001.bin