2001 July 1-September 30 AIPS Quarterly Report

AIPS Distribution - The 31DEC01 version of AIPS, updated daily, has now been distributed 387 times, to 325 sites, since December 2000. A total of 43 sites have retrieved multiple copies of 31DEC01, for an average of 2.5 retrievals each. Among the 73 registered users of 31DEC01 (most users fail to register), the number of machines is split almost evenly between Linux and various versions of Sun/Solaris operating systems, but 2/3 of the registrants have Linux as their primary architecture.

Key Developments -

• The AIPS group is working toward moving the midnight job and the master version of AIPS to NRAO-Socorro, since there are no longer any programmers in the AIPS group in Charlottesville. It is anticipated that these moves will be complete, and formally implemented, at the time of the frozen release of 31DEC01 in December 2001.
• In response to reports of anomalous source counts in the NRAO VLA Sky Survey, Eric Greisen has undertaken a review of the Gaussian fitting in AIPS, including a series of Monte Carlo runs to determine errors in fitted parameters for a variety of source models. Results will be presented in an AIPS memo, and are being presented at the ADASS (Astronomical Data Analysis and Software Systems) meeting in early October. The investigation also has resulted in changes to the fitting routines IMFIT and JMFIT as well as the combined source-finding/fitting task SAD.
• Data reduction utilities have been cleaned up in several instances. The VLBA utilities have had minor improvements made to permit use for spectral line data and to improve their use in polarization calibration. The VLAPROCS procedures (including VLACALIB) underwent a significant housecleaning in order to remove hidden inputs.
• The precession subroutines had several errors fixed to enable correct precession as well as interpolation between the input and output precession dates. Calculation of the relativistic light-bending and aberration have been modified. Correct matrices for inverse precession were developed and implemented.
• Testing of FRING and KRING has shown no significant differences between the two fringe-fitting programs for sources of high signal/noise; preliminary results show differences, but no systematic trends, in cases with low signal/noise or resolved sources. Tests are continuing, and are expected to be completed in the 4^th quarter.
• FILLM was modified to improve the handling of frequency and source parameters. In addition, the capability to load and use VLA weather data has been added.
• FITLD was modified to improve the handling of CQ tables, and to apply (or not apply) the correction for delay quantization in the correlator, depending on the correlator identity.
• UVPLT was modified to add the option of plotting quantities vs. the projected (u,v) distance in any position angle, not just along the axes.

Goals for Q4 2001 -

1. Continuing maintenance and user support.
2. Freeze and release 31DEC01 version, with primary copy and midnight job moved to Socorro.
3. Complete comparison of FRING and KRING for fringe-fitting under various conditions.
4. Develop and test task to separate primary and subreflector effects in VLBA holography.
5. Modify TECOR to permit use of ionosphere observations across UT midnight.
6. Develop a data-editing task to use the VLA weather information now provided by FILLM.