AIPS quarterly report, 30 June 1997 =================================== system ------ During the second quarter of 1997 we continued our emphasis on space VLBI related applications. Due to the delays in the launch, getting scientific data, and finally in finding fringes in these data, we decided to postpone the release of AIPS until July 1997. In the meantime, fringes have been found, and the first set of data has passed through the system. Though admittedly this is a very simple case, we consider the relative ease with which AIPS handled these data as encouraging. The midnight job is rapidly gaining popularity as a means to gain quick access to the latest AIPS developments. The number of non-NRAO sites that is running the AIPS midnight job has increased from six to nine. The three SGI workstations purchased by the VSOP/Halca project were installed and are experiencing heavy use. The four-processor SGI 'Origin 200' set a new AIPS mark record attaining a value of 14.1 for a single processor. This enormous power will be used exclusively for fringe-fitting and imaging space VLBI data. personnel --------- Athol Kemball left the AIPS group to join the AIPS++ group. A 12-month plan was issued in which we agree not to replace Kemball and concentrate on selected AIPS targets only. Two of these targets are porting of single dish applications from CVEX, and parallelizing selected pieces of code. Main aim of this plan is to guide the personnel transition from AIPS to AIPS++. It assumes that AIPS++ can take over major AIPS responsibilities one year from now; if this should not be the case, we will have to reconsider the AIPS staffing situation. software -------- 1 - AIPS failed on handling data taken with a 84-element Heliograph. A major rebuild using a larger maximum number of baselines fixed this particular problem. 2 - CLCAL now allows pass-through calibration, as well as 'proper' FQID and SOURCEID selection. By choosing 'CALP', CLCAL can now be made to pass CL table records for which no SN table record is found. This is dangerous but necessary if SN tables are constructed piecemeal as may be necessary for VSOP data. 3 - Dave Gordon of USNO has made a series of modifications to the Haystack Fringe (HF) table handling routines CL2HF and HF2SV. We have pushed these modifications through the Solaris compiler and have removed most non-standard FORTRAN statements. 4 - We added the ability of change the default printer from within AIPS. Previously, changing the default printer or file made it necessary to exit and re-start AIPS. 5 - The new task OBEDT allows editing data depending on values in the Orbit (OB) table. Another task, OBPLT, plots two columns in the Orbit table against each other. 6 - Tape access for the SGI machines failed, which prompted an update of several fundamental tape access routines. 7 - Space VLBI observations may require more than the 20 table types which is the current AIPS maximum. An effort was started to increase this number to 50. 8 - Increasingly, new operating systems support file sizes exceeding the 2.1 GByte limit. Work was initiated to add support of these file sizes to AIPS. Due to their inherently large sizes, this development is particularly useful for Space VLBI datasets. 8 - We further increased the support of parallelization by introducing the new verb PARALLEL. It allows the user to specify the number of processors that will be used for AIPS tasks that have been compiled with multiprocessor support. 9 - We modified IBLED to allow an overlay plot of model amplitude as derived from an external CLEAN component table. This plot can be selected or de-selected using a new "PLOT MODEL" option of the left-hand menu. The model plotting supports multi-field data, and is useful for editing gravitational lens datasets in particular. 10 - A new routine to fit polynomials was written, and was used in ACFIT to improve spectral baseline fitting. 11 - In the 'designated AIP' program we continued to lend support both to in-house and to external AIPS users. Gustaaf van Moorsel, 1 July 1997