Notes from Science Tools VRVS, November 22, 2006

[These notes are unedited and freely paraphrase the discussion; corrections are welcome.  SD]

J McEnery, P Nolan, T Stephens, T Burnett, A Cillis, J Carson, R Sambruna, S Digel

[This was to be a recap of the outcome of the GUC beta test plus further discussion of future directions for science tools development.  We had a 30-minute time slot sandwiched between other VRVS meetings.  No science tools development news was volunteered; the proximity to the GUC test and the short holiday week may have been contributing factors.  Owing to technical difficulties, Chris was unable to attend this VRVS meeting and we did not have a recap of the GUC outcome.]

SD - SSC coming to SC workshop?
JM - Agreement was that it would be useful to have SSC presence at the workshop but beta test pressure made that drop off the table
SD - VRVS


Science tools development issues


JM - Do we have plans to make the likelihood tool produce upper limits? E.g., for every burst we are going to want to say what the upper limits are.
JCar - Likelihood does report parameters of the model for each component regardless of how high the test statistic is.  So we do still have an estimated flux that might be considered an upper limit - maybe good enough for quicklook purposes
JM - I think that we need a strict upper limit - likelihood best fit probably does not err on the top side of the range

SD - What did GUC beta test users do for visualization?
JM - Counts maps, running likelihood, using gtbin to look at light curves; several people did pulsar analysis.  One person fiddled around with seeing whether SNR were extended sources, some looked at bursts.
JM - Overall people were left with a good impression

SD - Other comments?
JM - Something that did come up was the desire to get a set of independent components - each energy range independently fit like in catalog analysis - one can get that from existing tools but it is work.
JM - Comment on the accumulated livetime in FT1 file - essential for GRB analysis - but not necessarily since start of mission but wrt some reference time - otherwise not correct properly for deadtime for very bright transients.  Alternative would be to extract finely graded ft2 file for specific time intervals

SD - Will not meet next week owing to SC workshop.  Whether we meet the following week depends on coordination with the catalog group.