Date: Thu, 05 May 2005 19:46:17 -0400 From: "Mark M. Ito" To: marki@jlab.org Subject: skim problem notes, talk with Dave Talked to David Lawrence. The trigger supervisor roc keeps a time, in seconds, that has a fixed offset from absolute time between reboots, i. e., it starts at zero at each reboot. The code in the trigger package, to fill the eventid bank, tries get its time from this TS time, but to apply an offset to align it with unix time. It does this a begin run time, using begin run time. Problem is when one file at a time is analyzed, this is only the correct procedure for the first file. The idea of the eventid bank is to collect event-header-like information from many source, possibly applying standard corrections. To get the correct unix time, one needs to calibrate the ts roc time scale. Begin run time is the obvious point to do that, we do not in general (presently) have access to the ts roc time at begin run. This could be a calibration task.