The results of the PIBETA replays are on the dirac cluster,
 node15. The directory /sm_local/ contains two subdirectories:

 1. /sm_local/pb_dst7

 2. /sm_local/pb_dst7_toff

 The data and analysis that use the online timing are in
 first subdirectory. The data and analysis that use/will be 
 the best offline timing are/will be in the second subdirectory.
  
 The structure of the each subdirectory is as follows:

 beam subdirectory = pi_beam events
 chamber_reso subdirectory = MWPC resolution
 cosmics subdirectory = cosmics events
 dalitz subdirectory = Dalitz events
 dsc subdirectory = DSC analysis
 ggg subdirectory = SCX pi0->gg data and analysis
 kumac subdirectory = kumac macros 
 lfv subdirectory = lfv events
 michelsubdirectory = Michel events
 mueee subdirectory = eee (three MI) events
 mueg subdirectory  = mu->eg events
 mwpc subdirectory  = MWPC analysis
 pibeta subdirectory = pibeta events
 pienu subdirectory  = pi->enu events
 prompt subdirectory = prompt events
 random subdirectory = random events
 rm subdirectory = radiative Michel events
 rp subdirectory = radiative pion events
 rz subdirectory = HBOOK RZ files for all runs
 timing subdirectory = zero timing events
 tomo subdirectory = tomography events

 Each subdirectory above contains the *.rz file
 that is the data summary file for the reaction in
 question. The ascii data from which *.rz file is
 derived are ./data subdirectories. 

 The total size of /sm_local/pb_dst7 subdirectory is 
 at the moment 62G.

---------------------------------------------------------------

 The *.mid files skimmed for individual reactions are
 on the dirac cluster, node14. The top subdirectory is
 /sn_local/dsc. From there, the subdirectory structure is
 equivalent to the structure on pibeta@node15:/sm_local/pb_dst7
 that used the un-skimmed *.mid files for replay.

 beam subdirectory = pi_beam events
 chamber_reso subdirectory = MWPC resolution
 cosmics subdirectory = cosmics events
 dalitz subdirectory = Dalitz events
 dsc subdirectory = DSC analysis
 ggg subdirectory = SCX pi0->gg data and analysis
 kumac subdirectory = kumac macros 
 lfv subdirectory = lfv events
 michelsubdirectory = Michel events
 mueee subdirectory = eee (three MI) events
 mueg subdirectory  = mu->eg events
 mwpc subdirectory  = MWPC analysis
 pibeta subdirectory = pibeta events
 pienu subdirectory  = pi->enu events
 prompt subdirectory = prompt events
 random subdirectory = random events
 rm subdirectory = radiative Michel events
 rp subdirectory = radiative pion events
 rz subdirectory = HBOOK RZ files for all runs
 timing subdirectory = zero timing events
 tomo subdirectory = tomography events

 Each subdirectory above contains ./*99_mid, /*00_mid, and
 /*01_mid subdirectories containing the skimmed midas
 files for years 99, 00, 01 for the reaction in question, eg
 pibeta@node14:/sn_local/dsc/pibeta/pb99_mid, etc

 We have to decide which portions of the above structure
 we want to back up to magmetic tapes and/or DVD disks.
 The kumac macros and fortran analysis files should be
 probably back up on one media and the HBOOK data summary
 files for individual reactions on the separate media.
 That way we can distribute the DST to the collaborators more
 easily.

---------------------------------------------------------------

 All odb files are on dirac, node14, at

 /sm_local/odb1999
 /sm_local/odb2000
 /sm_local/odb2001

---------------------------------------------------------------