Science Tools Working Group
We did not meet this week but most likely will meet on June 28.
The current version of the Science Tools remains v7r3
Databases and related utilities
No news.
Likelihood analysis
From Jim: "I fixed a couple bugs in the FileFunction model that allows one to use an ascii file with columns of energy and differential photon flux as the spectral model for a source."
"I have also determined that the biases that are affecting Likelihood fits to data generated using the DC2 IRFs arise almost entirely from the fact that the DC2 energy dispersion functions are not centered on or symmetric about the true energy of the photon and tend to result in systematically lower measured energies on average." (See slide 7 of my DC2 Kickoff talk on IRFs.) This effect should have been obvious in retrospect. I am working on a remedy for this."
Routinely in likelihood analysis we ignore the energy dispersion, because the response functions do not depend very strongly on energy at the level of the energy resolution. However, because the energy dispersion has an asymmetric distribution, Jim has realized that we are effectively introducing a systematic bias to lower energies in selecting the response functions to use in evaluating the likelihood function.
GRB tools
No news.
Pulsar tools
Masa and James worked on (or 'refactored') the TimeRep class that will implement the conversions between representations of time - like MET, MJD, etc., with a goal of having the class complete and in use by gtpulsardb, gtephcomp, gtpsearch, gtpphase, and gtophase by the end of the current build (July 21).
Observation simulation
No news
User interface and infrastructure
James and Analia have implemented some changes to st_graph to support plotting from gtlikelihood . James is continuing to work on ape, mainly bug fixes, for public release as part of HEAdas. Integration with the science tools will follow.
Source Catalog
The catalog group met this week, if you are interested. Jean presented initial results from evaulating significances for the 'true' sources in DC2 and the effects of removing the background from the data set. Also, he gave a detailed presentation of an updated estimate of the number of CPU days required for a catalog analysis for the current pipeline. Follow the link above.
Science Tools Working Group
We did not meet this week and will not meet next week either. Best guess is June 28 for the next Science Tools meeting.
The current version of the Science Tools is v7r3; Jim tagged this last week, at the end of build cycle 18. 'DC2 installers' for this version are not (yet) available. Navid has not been asked to make them, and I think that we may want to be selective in which versions we ask him to build, at least until the system for making distributions for the various supported platforms is automated. Here's what's new since v7r2; the ScienceTools package release.notes have a concise summary of the new features.
Databases and related utilities
No news.
Likelihood analysis
From Jim: "I re-activated the counts spectrum plotting option for gtlikelihood and discussed further development of the option with James Peachey and Analia Cillis; I fixed a bug in the xml output for the FileFunction source in Likelihood, and modified the interface in the optimizers package so that the C++ version of the Minuit libraries that are shipped with ROOT can be used by Likelihood."
GRB tools
No news.
Pulsar tools
Masa and James are working on the TimeRep class that will implement the conversions between representations of time - like MET, MJD, etc.
Observation simulation
Jim has reimplemented the FileSpectrum source - which allows an arbitrary spectrum for a point source to be specified as an ASCII table. The concept has not changed, but the implementation is more robust.
Max has enhanced PulsarSpectrum (to v1r2) to allow an arbitrary phase dependence for pulsar spectra. This is an important advance. Right now, I don't know how the user specifies the inputs, but the new version appears to be part of v7r3 of the ScienceTools, so it is certainly available for playing with. Max reports that a small update to v1r2p0 is in the works.
User interface and infrastructure
No new news. James reports that he is doing some more work on APE, the replacement for PIL in HEADAS. I have not worked myself up to asking what APE stands for, partially out of fear that it stands for A PIL Enhancement.
Source Catalog
The catalog group did not meet this week.
Science Tools Working Group
We did not meet this week. The next meeting will likely occur June 14.
DC2 Fallout
DC2 was a huge success (as expected). Participants made extensive use of virtually all of the ScienceTools and provided much feed backfor developers to chew on.
Databases and related utilities.
No news.
Likelihood analysis.
JC met with Chris Shrader, Dave Davis, David Band, James Peachey and Analia Cillis to chart the course for developmentof the Likelihood tools in the coming months. Most of the proposed enhancements are in response to user requests that arose during DC2. Highlights include integrating of a plotting option in gtlikelihood, access to model flux values from gtlikelihood, more robust error calculations, confidence contours for fit parameters, and a generating a warning if the spectral model does not fit the data well (but not providing a goodness-of-fit measure). Analia Cillis will be helping out in the implementation of some of these features.
GRB tools.
No news.
Pulsar tools.
From the GSSC weekly report: "Masa Hirayama and James Peachey worked on the infrastructure of the pulsar analysis tools, in order to enhance users' convenience, including automated time conversions between time systems."
Observation simulation.
No news.
User interface and infrastructure.
James Peachey reports that the replacement for PIL, called APE, is finished, having been included in a recent HEADAS release, but he must update the hoops interface before we can use it with our software.
Source Catalog.
The catalog group met yesterday. Topics included a recap of the DC2 presentations, a scheme for comparing image-based source detection methods, generating a reference list of sources used for DC2, including appropriate spectral models, and a presentation by Larry Wai on detecting extended sources at the 5-sigma level.