Science Tools Working Group
We met this week. Our next meeting will be on a to-be-determined date after the GLAST Symposium.
The current version of ScienceTools remains v7r6p1.
Data products: No news
Databases and related utilities
Not much news. Richard has posted the core of a proposal for how the data should be organized for access; this is open for comment
Likelihood analysis
No news; Jim has a couple of bug reports from Jean, but work on them has not started yet.
We had a long discussion about the state of our understanding about how Test Statistic values relate to significances. Short answer: incomplete. Some potentially-useful directions are being explored - like simulations that Pat is developing to test a potential analytical form being worked out by Ramani Pilla (Case Western Univ.) for the distribution of TS.
GRB tools
No news
Pulsar tools
Masa and James are continuing to work on implementing the blind search for peridicity; they are also working out numerically stable ways of calculating the loss of significance for periodicity detection due to trials factors.
Observation simulation
Max R. is continuing to work on PulsarSpectrum to simulate binary pulsars and has not had a chance yet to try to reproduce the apparent bug mentioned last week.
Toby has got a 'moon' point source that is fixed in the reference frame of the moon working in Gleam simulations. Next step in terms of fidelity would be to implement orbital parallax.
Toby has modified the flux package to allow an offset to be specified for the assignment of MC_SRC_ID values and Jim has implemented this as a hidden parameter for gtobssim. This allows some flexibility for creating distinct MC_SRC_ID values in models that are generated as separate runs for different components of the sky model.
User interface and infrastructure
No news
Source Catalog
Did not meet this week.
Science Tools Working Group
We did not meet this week, owing to the GLAST science promotion event at Goddard on the 17th; no URL was available at press time. We will meet next week, and the next Catalog VRVS meeting will be on the 31st.
The current version of ScienceTools remains v7r6p1.
Data products: No news
Databases and related utilities
No news. Tony has asked for details about how the various kinds of data that the Data Catalog/Data Server will keep track of will be accessed and how these data types will relate to each other. Richard, Julie, and I are still really just about to draft something for comment.
Likelihood analysis
No news
GRB tools
No news
Pulsar tools
From James: "James and Masa continued work on the pulsar tools. While refactoring the periodSearch package in order to add support for the A4 tool (the limited blind search FFT tool), they discovered problems with the way the chance probability is computed, specifically that the number of independent trials was not being taken into account. They are partway through correcting that problem and will then continue with the scheduled refactoring and work to support the new FFT tool."
Observation simulation
Max R. has updated PulsarSpectrum to simulate binary pulsars (i.e., pulsars in binary systems - many more timing parameters); I do not know how extensively it has been tested. Unrelated to this update, recent versions of PulsarSpectrum (more recent than in v7r6p1 of the ScienceTools) report an out-of-time-range error for pulsars that could be simulated with v7r6p1. Max is looking into why.
User interface and infrastructure
Jean's JIRA issue about discovering that most of the output from gtlikelihood does not go to stdout has been resolved thanks to James' modifications to st_stream.
Source Catalog
Did not meet this week.
Science Tools Working Group
We met last week but we might not meet next week, owing to a GLAST science promotion event at Goddard (primarily involving local universities, I think) on the 17th.
The current version of ScienceTools remains v7r6p1.
Data products: We discussed the status of the LAT science data products (the ones that get delivered to the GSSC and hence are part of the 'interface') at the last science tools meeting. The open issues range from important details (like event numbers and diffuse response columns in FT1, or 'deadtime' in FT2) to whether a previously-planned data product like 'configuration history' can be made relevant and useful. A working group will be spawned with GSSC & LAT representation in advance of configuration control.
Databases and related utilities
No news. Tony has asked for details about how the various kinds of data that the Data Catalog/Data Server will keep track of will be accessed and how these data types will relate to each other. Richard, Julie, and I are [really just about to] draft something for comment.
Likelihood analysis
Jim corrected the calculation of the predicted number of events from diffuse sources; the error affected only plotting of results, not the optimization of the likelihood function itself. Jim has also modified the handling of names of diffuse sources so that the underscore character can now be part of the name without confusing the interpretation of diffuse response columns in FT1 files.
GRB tools
No news.
Pulsar tools
James reports that he and Masa are continuing to work on the tool for blind searches for pulsations. "We're integrating it into the periodSearch package, and we've been doing some refactoring in order to add the new code without duplicating old code."
Observation simulation
Jim has fixed (but not yet extensively tested) a problem with using log-parabola source spectra in simulations - the simulations were using a lot of memory.
Richard has a new source for simulating microquasars, which he'll describe at a forthcoming science tools meeting. Max R. has updated PulsarSpectrum recently, but I do not have any details.
User interface and infrastructure
You may have noticed Jean's JIRA issue about discovering that most of the output from gtlikelihood does not go to stdout. Jim pointed out that this was due to aspects of the implementation of st_stream. From James: 'I wrote to Jim about that and we agreed on a simple solution, which I'll check in pretty soon. There's a stream for "information" which is what Jim uses. That stream currently is redirected to the standard log stream (which is stderr). I will direct that instead to stdout and, as Jim suggests, add a new method to return a stream for logging.'
Source Catalog
Met this week. Problems with the SLAC reflector prevented me from personally participating. Of possible interest to this meeting is Jean's report on running the catalog analysis pipeline for the SC1 (1-year data). How many sources did it detect? What problems are they facing? Also Paolo or Gino presented a method for identifying regions of the sky that have variable sources.