;----------------------------------------------------------------------- ; Copyright (C) 2009-2011 ; Associated Universities, Inc. Washington DC, USA. ; ; This program is free software; you can redistribute it and/or ; modify it under the terms of the GNU General Public License as ; published by the Free Software Foundation; either version 2 of ; the License, or (at your option) any later version. ; ; This program is distributed in the hope that it will be useful, ; but WITHOUT ANY WARRANTY; without even the implied warranty of ; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ; GNU General Public License for more details. ; ; You should have received a copy of the GNU General Public ; License along with this program; if not, write to the Free ; Software Foundation, Inc., 675 Massachusetts Ave, Cambridge, ; MA 02139, USA. ; ; Correspondence concerning AIPS should be addressed as follows: ; Internet email: aipsmail@nrao.edu. ; Postal address: AIPS Project Office ; National Radio Astronomy Observatory ; 520 Edgemont Road ; Charlottesville, VA 22903-2475 USA ----------------------------------------------------------------------- This file must be used to record any changes made to the Master version of AIPS kept in SOCORRO. An entry is: line 1: Date Program name Your name line 2: What was done (may span multiple lines) line n-1: Where has this change been moved (e.g. 15JUL94, nowhere) line n: Blank The details of line 1 are important since programs are used to find specified strings and do conversions on them. It MUST be typed as follows: 1 blank, 4 digits, a period, two blanks, then the Month, day, comma, year, some space, a title, some more space, your name. *********************************************************************** *********************************************************************** Changes while 31DEC10 is TST *********************************************************************** *********************************************************************** 13023. December 2, 2009 version Eric Changed 31DEC09 to 31DEC10 in: GETRLS.FOR INSTEP1 install.pl update.pl ZXHLP2.PL UPDCONTROL (comment) AIPSPATH.* Changed LIBR.DAT in $SYSUNIX, $SYSSUL, $SYSMACIN, $SYSMAXPP to put Z routines in APLSUB library. Moved nowhere. 13024. December 2, 2009 UVRFI.FOR Leonia This new task mitigates RFI using the two algorithm: 'CEXP': Fourier transform of the data at each time interval, each baseline, each channel, and following complex Hogbom clean subtracting one or more components from the "dirty map", using precalculated "dirty beam". The cleaned complex value at zero fringe rate is the restored visibility value. 'CIRC': Fitting the circle at the complex visibility plane. This agorithm is a modernized algorithm of Raman Athrea ApJ 696, 885, 2009, May 1, which includes fitting of the circle radius increment. Some flagging of the data at channel by channel is implemented. Data must be 'BT' order!!!!!!!. Moved nowhere. 13025. December 3, 2009 CookBook Eric 1 : Updated numbers, aips++ remarks, misc wordings 2 : Drop TIMDEST reference, misc minor wordings 3 : RECAT, AN file format, drop HORUS, MX, TIMDEST, WHATSNEW, add FILEZAP, correct VLA archive web address 4 : 3C147 model, RLCOR, ATLOD now AIPS, SNFLG flags data, GETJY robust average 5 : IMAGR (DO3D false, autobox, IM2P(7), restart after filter, OVRSWTCH, baseline-length dependent averaging) MX/HORUS, SABOX, FILIT, UVFND phase, FQUBE 6 : ISPEC/BLSUM make SL files, REROAM amd ROAMOFF, FILIT, CAPLT, UVPLT NCHAVG option 7 : ISPEC, BLSUM slices; UVMOD text file, FQUBE, FARS 8 : ISPEC, BLSUM slices; MX, HORUS gone; IMAGR autobox/carry boxes over between channels; FQUBE 9 : SNPLT DOBLANK, RLCOR, FITLD read pre-processed archive data but without DOCONCAT, VPLOT max/min, CLCOR 'IONO', CAPLT, FRING DPARM(9) no rates, SNCOP for copying SN tables, BLCAL ICHANSEL and BLCHN 11 : Drop TIMDEST several places 12 : Only private CA, FILEZAP, FQUBE, DELAY, several misc. 13 : Updated all changed ABOUT files A : Drop MX, HORUS, WFCLN, MAPIT proc; reword VLACALIB. B : Wording corrections I : update J : update Recipe: rearranged only LaCook.PL revised to correct version(s). Moved PS to shared TEXT, $DOCTXT/LaCook* to 31DEC09 and 31DEC10. 13026. December 3, 2009 ANCHECK Eric Effelsberg is EB not EF - corrected ANFIND.FOR and ANCHECK.001 in NEW and TST and ANFIX8 in the patch page for 31DEC08 (OLD). Both EF and EB are now recognized per Craig. Moved as described above. 13027. December 4, 2009 UVRFI.FOR Leonia The subroutine name CEXP is substituted by CMPEXP, because CEXP has been used already. Moved nowhere. 13028. December 7, 2009 SPLIT, SPLAT Eric SPLIT and SPLAT assumed that the header value of the alternate reference pixel was correct even for multi-source files. This is not necessarily the case and should always be the frequency reference pixel. Changed usage. Moved to 31DEC09 this date. 13029. December8, 2009 ATLOD Eric Messed up one of the 2 calls to TABAN. Moved to 31DEC09 this date. 13030. December 10, 2009 TIMDEST Eric Removed messages that mentioned TIMDEST - it no longer is available to users in PRTMSG, AIPSC (added block), and ZDCHIN. Moved to 31DEC09. 13031. December 10, 2009 COMB Eric Added code to copy tables to the output except for CC tables. Moved to 31DEC09. 13032. December 11, 2009 install.pl Eric Added code to trap the current list of hosts that are found in a pre-existing HOSTS.LIST file and the to reread that file to make sure that those hosts have the current $ARCH and $SITE that install.pl assumes. This should stop mysterious errors which arise from a mis-match of the old and new values of these critcal strings. Moved to 31DEC09 December 29, 2009. 13033. December 14, 2009 UVRFI Leonia Changed the threshold 2.5 to 1.5 to allow the substruction at the Fourier tranform (opcode='cexp') more close to the zero fringe rate. Add small change to the help file Moved nowhere 13034. December 14, 2009 RM Eric Changed the form of the input data to have an FQID axis as produced by MCUBE and FQUBE. Added another method due to F Zhou. Moved nowhere. 13035. December 14, 2009 EVASN, EVAUV Eric NEW tasks written to evaluate pipeline results. EVASN looks at statistics of the SN table (amp rms, phase stability) and EVAUV looks at image statistics plus the uv statistics after a model subtraction and division. They are under development and may need more tetsts and certainly need to return some parameters to the pipeline. Moved nowhere. 13036. December 14, 2009 SAD Eric Added message saying that an RMS image is being used. Moved nowhere. 13037. December 15, 2009 Memo 114 Eric Put back a revision of the memo - numerous minor errors plus BITPIX, 3 types of weights, more care with EQUINOX and EPOCH, antenna diameter and full-width at half-maximum, units corrected for dispersion, etc. Corrected web pages. Moved nowhere (all versions get it). 13038. December 18, 2009 UVRFI Leonia Factor at comparison the channel amplitude with mean amplitute through all channels has been constant (1.5). Now it is the input parameter APARM(6). The help file is changed relevantly. Moved nowhere. 13039. December 18, 2009 SMOOTH Eric SMOSP made an erroneous reference to CATBLK (output header) rather than CATUV (input header). In IBLED it got a very wrong set of increments through the data - so IBLED messed up badly if SMOOTH was set. Most tasks have not messed up the CATBLK enough to matter after the INIT call to UVGET but before the READ calls. Discovered that spectral smoothing was occuring before delay calibration application in the before bandpass smooths. This is a very bad idea but not important for small delays. Moved to 31DEC09 this date - these have been with us a long time. 13040. December 21, 2009 EVASN Eric Corrected the robust average to handle the case where RMS is zero - samples are included in the next average when they are less or equal factor * rms. Before it included only less than and that excludes all when rms is 0 (e.g. amplitude on a phase only SN table). Oddly, the rms changed in other circumstances too. Moved nowhere. 13041. December 23, 2009 BLCAL Eric The help file referred to writing an output file. BLCAL does not do this. Dropped the remark and cleaned up the Explain typing. Fixed history file writing in the Fortran since it left out many of the calibration parameters and was coded oddly. Moved to 31DEC08 this date. 13042. December 23, 2009 BLCHN Eric Dropped remark in help file - output is written whether or not there is a division by the source model. Changed Fortran to compute spectral rms and honor the print option. Changed help and Fortran to offer a post-solution spectral smoothing identical to the usual smooth. Changed Fortran to write all calibration parameters to the history file and also cleaned up the rest of the HI writing. Moved to 31DEC08 this date. 13043. December 23, 2009 CVEL Eric It processes only one FREQID - changed help to emphasize this. Fixed long line in LISTR.HLP. Moved nowhere. 13044. December 25, 2009 TVFLG Eric Tried to read the number of records in an antenna file equal to the highest antenna number in that file. That is wrong. Changed the length finding routine to issue a warning if some possible antennas have no position info and so have length 0. Moved to 31DEC09 this date. 13045. December 25, 2009 SETJY Eric It issued a message about available models for each IF and had messages saaying model may be available about some of the bands. Changed to issue only one message per source and make all certain about having models. Moved to 31DEC09 this date. 13046. December 25, 2009 ISPEC Eric 1. Fixed return quickly to return quicly on DOCRT 0. 2. Fixed setting of status so it will clear at end of making plot file. Moved to 31DEC09 this date. 13047. December 25, 2009 IMEAN Eric Counted histograms as real*4 - limit of accuracy is 2 or 4 million after which count ceases. Changed to double precision. Moved to 31DEC09 this date. 13048. December 25, 2009 MCUBE Eric Found some code that limited the axis on which we are building to 2048. Fixed. Moved to 31DEC09 this date. 13049. December 29, 2009 CALIB, FRING Eric The handling of the index table, dividing the scan into equal increments ignored the start and stop times. Fixed to change the scan length and start and stop times to honor the overall start and stop times (from TIMERANG). Moved to 31DEC09 this date. 13050. December 30, 2009 LPOPEN Eric For batch, it set DOCRT = -1. This should be MIN (-1, DOCRT) since DOCRT = -2, -3 etc have meanings. Moved nowhere. 13051. December 31, 2009 FRING Eric Add CHINC option - channels are averaged before the fitting. Moved nowhere. 13052. December 31, 2009 ZMEMRY Eric ZMEMRY took an argument of number of words wanted but passed number of bytes to ZMEMR2. Changed to pass number kilobytes to ZMEMR2 and changed that to use a "long" for number of bytes. FRING is now very dangerous. Moved nowhere. 13053. January 1, 2010 SAD, JMFIT, IMFIT Eric Added DOCRT=-4 option to do a special format print to a text file with one line per source component. The format lists file name, channel, peak flux with error, offsets with errors, reference coordinate, Gaussian size and position angle with errors, Clean beam size and position angle, image rms, and component number. Angles are in arc seconds or milli arc seconds depending on the cell size. The scale used appears in a header line. Fluxes and noise are also converted to milli or micro Jy with a code displayed. Header lines only appear in new output files. Moved nowhere. 13054. January 1, 2010 AIPS Eric Changed copyright date to 2010 Moved nowhere. 13055. January 1, 2010 SUFIX Eric Changed it to allow the user to specify the name, calcode, and position of an existing source and change the wrong source numbers to those of that existing source. Moved nowhere. 13056. January 5, 2010 MFPRT Eric Forbad printing to the line printer - allowed files or the CRT only. Dropped that limit. Moved nowhere. 13057. January 5, 2010 IMSTAT, IMEAN Eric Images exceeding 2^31 - 1 pixels (e.g. 400x400x16112 in my case) will be upon us soon. Changed counters in AU6D and IMEAN to double precision float and modified formats. Moved to patch January 18. 13058. January 6, 2010 COPTAB Eric COPTAB copies tables as appropriate from one uv file being calibrated possiby to the output data set. For CL tables, it should write a null CL table 1 when DOCAL is true. Changed it to do this, calling CLNULL. Got CLNULL from SPLAT to make it a separate subroutine (it is similar to CLSEL) and changed AVSPC local COPTAB similarly. Previously CL table 1 was copied but it can contain some a priori calibration info. Moved nowhere. 13059. January 8, 2010 VLARUN Lorant Modified procedure numtab to make it safe against long keywords Moved nowhere. 13060. January 11, 2010 SPFLG Eric Change format of "Grid antennas" to be "Grid Baselines" to show what is actually happening. Moved nowhere. 13061. January 11, 2010 ISPEC Eric Correct it to determine the max/min default on the smoothed spectrum when smoothing is specified. Added XYRATIO adverb to allow plotting a rectangular area more suitable for large numbers of spectral channels. Moved nowhere. 13062. January 11, 2010 EXTLIST Eric Updated AU8A for changes in ISPEC adverbs. Add XYRATIO to display. Moved nowhere. 13063. January 13, 2010 JMFIT Eric Fixed handling of DOCRT which had gotten messed up. Moved nowhere. 13064. January 13, 2010 TVMOVIE Eric Changed AU5D and the help to allow LTYPE to have all normal values. They are converted to values from 1-6. Moved nowhere. 13065. January 13, 2009 TVANOT Eric Changed AU5B and help to use COMMENT rather than KEYSTR to get up to 64 characters rather than only 16 max. Moved nowhere. 13066. January 13, 2009 ZXHLP2 Eric Changed ZXHLP2 to check if $AIPS_BROWSER is set and, if not, to set it to firefox. Added info to XHELP help file. It produces an error message but works now. Moved nowhere. 13067. January 13, 2010 ELINT Leonia ELINT has calculated argument of an array adding LONGINTEGER and regular integer sending the sum to the regular integer. It should be sent to the longinteger variable. Having fixed this ELINT works at a 64 bit mashine as well as at a 32 bit mashine. Moved nowhere. 13068. January 13, 2009 LWPLA Eric Changed formats inside to allow for sizes > 33 inches. Also changed to allow larger fonts. Moved nowhere. 13069. January 15, 2010 FITTP Eric The DOSTOKES option (convert to I,Q,U,V) is just too dangerous. It can be set accidentally and then data may be saved without calibration being applied (so not a real Stokes at all). Drooped it from FITTP Fortran, WRTPROCS RUN file, and help files for DOSTOKES, FITTP, FITDISK, and WRTDISK. The DONEWTAB option is not needed and was also dropped. Moved nowhere. 13070. January 18, 2009 DOBAND Eric Corrected BPASET, restoring ISVLA variable. DATBND uses it to decide between all channels and the central 3/4 and we do not want to use the outer channels except for crazy VLBIers who insist on it. Moved to patch this date. 13071. January 21, 2010 IMFIT, JMFIT, SAD Eric These tasks on DOCRT=-4 had two problems: (1) The Clean beam parameters were those of the image in pixels and so were not corrected for rotation and for odd position angles used in SAD. (2) When the noise in the peak flux was larger than 1/10 of the peak, the format could run together for the peak and the noise. Added an invariant header Clean beam parameter to avoid the first problem. For the second added tests on the error bar level 10 times tighter than those on the peak so that F7.3 will always make a blank in the first character. Also made the test consider the nead for a minus sign on negative fluxes, forcing different tests for negative than positive fluxes. Moved nowhere. 13072. January 21, 2010 INDXR Eric Changed it to treat the EVLA as the VLA. This means it will want a weather table. Changed it to accept that a missing weather table means use 100% seasonal model with no surface weather contribution to the estimated opacity. Moved nowhere. 13073. January 21, 2010 PRTUV Eric Corrected program to set the end time to infinity if it is specified <= start time. Enhanced testing on time to quit when the sort order is T* and the end time is exceeded. Corrected testing on antenna numbers which had a typo! Moved nowhere. 13074. January 22, 2009 PRTAN, FILLM Eric FILLM: Rename pads to have 2 digits always including leading 0 for the first 9 pads. PRTAN: Made it handle the new form of name, converting old forms to it. Improved the VLA map to include missing stations on the North arm as well as on the other 2. Moved nowhere. 13075. January 23, 2010 LOCIT Eric LOCIT actually computes corrections to Bx and By at the longitude of the antenna. Added change to convert these corrections to be with respect to the longitude of the array. For the EVLA, this is at Greenwich, whilst for the VLA it was in the center of the Plains of Saint Agustin very near the antennas. Other changes: 1. Changed pad names like W4 to be W04 2. Dropped the .OPR and .SUM output files 3. Changed it to read a pre-existing .VLA file acquiring any coordinate changes contained therein. Then it renames the .VLA and writes a new one with the new sum of changes. 4. Added a .001 file with a CLCOR script containing the changes. 5. Added a .PAR file for the PARMENATOR with the sum of changes. 6. Made another plot file based on BPARM(9) request, plotting the X, Y, Z offsets along each arm in 3 plots (one page). Moved nowhere. 13076. January 25, 2009 FITLD Eric Olaf Wucknitz (Bonn Uni) reported another typo causing the polarization type stored in the AN table to contain garbage. Moved to patch this date. 13077. January 25, 2009 UVFND Eric Changed default APARM(1) to 1.E8 in all cases for CLIP. Added OPCODE 'WTUV' => do CLIP and check reasonable weights with WTUV. Otherwise, set WTUV = -1 and ignore weights. Moved nowhere. 13078. January 26, 2009 FARS Leonia The option of CLEAN at the Fourier transform is added now by new inputs: GAIN and FLUX. The old version subtracted the Fourier component at maximum amplititude by hundred percent of its value. The new version subtracts a portion of the maximum amplitude component under control of GAIN. The subtraction is finished when the iteration number achieves NITER or the clean flux achieves FLUX. Moved nowhere. 13079. January 28, 2009 CLCOR Eric A variety of places did not check to see if all pointers to the CL record were valid. It was then possible to write in RECORD(0) which damaged the SIN of the declination carried around in COMMON. The symptom was an error in shifting phases for position shifts that were large enough - or scans that lasted long enough perhaps - when there is only one polarization. When there are two polarizations all pointers are valid. The error appeared in the code in March 2006. Moved nowhere - lets wait and then move to patch. 13080. February 2, 2010 HI lines Eric Added HI lines to CLCOR ATMO et al input file name UVCOP Flag table usage TECOR Changed to make INPUTS object know actual CL versions so they will come out in HI FITTP DATAOUT or OUTTAPE to HI FITAB DATAOUT or OUTTAPE to HI Moved nowhere. 13081. February 4, 2010 UVGET Eric Changed UVGET so that EVLA data have DOWTCL forced to false. Changes all tasks with adverb DOCAL. Moved nowhere. 13082. February 4, 2010 WAITTASK Eric Changed AU2 to allow WAITTASK on AIPSC - just not on AIPS and AIPSB. TPUT and VPUT are still disallowed for all 3. Moved nowhere. 13083. February 5, 2010 CLCOR Leonia OPCODE 'CLOC' has not modified RATE column of CL table. So CLCOR is modified to do it. Moved nowhere. 13084. February 5, 2009 IMEAN, IMSTAT Eric The format printing the number of pixels has to be corrected for a previous format that shifted the decimal place. The number printed was a factor of 10 off. Moved to patch this date. 13085. February 10, 2010 Delay correction Eric Amplitude should not be corrected for a delay error unless we know how much averaging was done after the initial correlation. Commenting out the current correction. Moved to patch this date. 13086. February 11, 2010 APCAL Eric The read routine used the loop counter as a row number argument to TABWX. But that routine increments the row counter so only every other row was read and occasionally it would try to read off the end. Moved to patch this date. 13087. February 12, 2010 Remote batch Eric Changed POPSDAT Added REMQUE adverb for SUBMIT. BATQ Added HOST argument, retyped the code in modern style, if HOST not blank, open file via NET0:HOST/ and the trailing characters from ZPHFIL. LPOPEN Fixed call sequence to BATQ - the HOST is blank here since we are asking what is happening in this batch job. AUB Added adverb REMHOST for UNQUE, QUEUES, JOBLIST. Correct error handling, file closing. AUA Add REMHOST to send to AIPSC. UNQUE Help file with added adverb QUEUES Help file with added adverb JOBLIST Help file with added adverb SUBMIT Help file with added REMHOST, REMQUE AIPSC Help file with added REMHOST AIPSC Add REMHOST adverb. Change where output files are, add code to find NINTRN and NBATQS parameters for the remote machine, add code to invoke START_QMNGR on the remote machine. Changed call seq to BATQ START_QMNGR New script needs to be in AIPS_ROOT QMNGREXEC New script executed by START_QMNGR ZSTRTQ New $AIPGUNIX program to start QMNGRx install.pl Added START_QMNGR to list to copy to AIPS_$ROOT AIPSROOT.DEFINE Added START_QMNGR to list. UPDUPDATE MNJ to copy START_QMNGR to $AIPS_ROOT Moved nowhere. 13088. February 12, 2010 ADDDISK, REMDISK Eric Created verbs to add a machine's disk to the list of available disk within AIPS and to remove same. Changed POPSDAT Add verbs START_AIPS To create a logical DADEVS_FILE with the dadevs file we are using this time DISK.INC New include to hold arrays to track added disks AIPS Initialize new arrays to keep track of disks added AU3B Add code for verbs - check and call DSKASS DSKASS New subroutine - reads the dadevs file and does the assignments on ADDDISK and uses Common arrays to remove disks. GNETSP New subroutine to find a list of disks in NETSP to find the timedest and user number restrictions which apply. ADDDISK New help file. REMDISK New help file. Moved nowhere. 13089. February 12, 2010 VLANT Eric Add a JTRIM - some machines seem unable to read text files correctly every time. Moved nowhere. 13090. February 12, 2009 FITTP, FITAB Eric Found a HI write line where DATAOUT > 33 characters caused a format overflow. Corrected it - this is a new bug (Feb 2). Moved nowhere. 13091. February 14, 2010 SETJY Eric Put in Rick Perley's latest and greatest values for the standard calibration sources. Seems to change things by < or about 1 percent. Changed APARM(2) values, shifting everything up by 1 and making 1 be what SETJY has done by default for a long time (21 Dec 2000). Moved nowhere. 13092. March 1, 2010 CVEL Eric My recent code contained the ability to go into an infinite loop when minor roundoff errors with times arose. Added a slight extra bit to the time and added other checks to avoid the infinite loop. BUT - in the course of testing, I found that with single-source files, the source name was null which prevented the message showing the shift or lack thereof from being completed properly. With single-source files having an index (NX) table, the task only shifted the first scan! This fact was obscured by the error in the source name. Corrected. Moved to patch this date. The lack of shift is old and maybe needs to be in the NRAO newsletter. 13093. March 3, 2010 LPSEC Leonia I have updated LPSEC because the new leap second occured (Jan1, 2009) since the last update (Jan1, 2006). The Julian days were given (in the table) at the format corresponded to REAL, although it must be DOUBLE PRECISION. I have fixed this. Moved nowhere. 13094. March 3, 2010 Pol cal Eric The array to apply polarization cal is dimensioned Number of baselines times 16 (4x4 array) times Number of IFs. Changed PUVD.INC parameter XPTBSZ from 16384 to 196608 to accomodate the EVLA. It should be 5767200 to accomodate 90 antennas and 90 IFs. Moved nowhere. 13095. March 3, 2010 DATUTC Eric Changed UVFIX to zero DATUTC if TIMSYS = 'UTC' and clarified the meaning of UVFIXPRM(12) in the HELP to be time system used by data - UTC rather than assuming IAT is in the data. The leap second routine is used only for array 'VLA' and so is not used for EVLA (which is what we want!). Changed ANINI to fill in IATUTC correctly from RDATE and subroutine LPSEC and on read to use IATUTC for DATUTC only if TIMSYS not UTC. Moved LPSEC from $APLNOT to $APLSUB. Moved nowhere. 13096. March 3, 2010 UVLOD, FITLD Eric Added code to re-write the source table with AIPS computing the apparent coordinates for all sources. The apparent coordinates received from various places are not filled in or are filled in incorrectly. NOte that these are used only a little and so it does not matter exactly what time to which the coordinates apply. Moved nowhere. 13097. March 4, 2010 GAL Eric The output of the fit parameters neglected to include the Brandt index and used formats that are potentially too small. The HCN routine which computes the residual image from input text, e.g. varying models in rings, was a bad mess. It got the geometry right, but did not compute any of the models that were in the rest of the program. Changed it to have sensible defaults for DPARM(1) and DPARM(2), to avoid zero divides, and to honor the desired model type. Moved nowhere. 13098. March 4, 2010 UVFLG Amy Flagging on shadowing was not working because UVFLG was not calculating the baseline length correctly. This was traced to the fact that ARRTYP had not been taken out of the call sequence for the subroutine BASLIN. Eric said he took that variable out of the BASLIN last Sept., so shadowing in UVFLG has been broken since then. Moved to patch March 5. 13099. March 4, 2010 ANTAB Eric The KEIN value buffer allowed only 100000 values. Changed that to 5000000. Moved nowhere. 13100. March 5, 2010 FITLD Eric RDATE is not a required keyword. Fixed tests on it to check only non-blank values. Moved to 31DEC09 since it alarms EVN users. 13101. March 5, 2010 Remote batch Eric Changes: QMNGREXEC Must be executable - tweaked system to try to force this BATER Procedure in $SYSUNIX did not set print devices, disks, TVALT, etc. Fixed it up. BATER Changed Fortran stand-alone program to do new stuff with REMHOST, REMQUE, etc. AIPSC Fixed to open input from BATQ rather then REMQ. Moved nowhere. 13102. March 5, 2010 ADDDISK, REMDISK Eric Changed DSKASS to avoid having a disk which failed to get assigned remain in the list. This prevents it from being deassigned or reassigned. Moved nowhere. 13103. March 5, 2010 FRPLT Eric Changed it to interpolate across empty time bins rather than FFT with bins at 0.0. Time increment edges can cause a regular pattern of empty bins which simulates a period within the data that is not present. Moved nowhere. 13104. March 9, 2010 BLSUM Eric The task did not consider the possiblility that the target and blotch images would have different Clean beam sizes. Corrected this. The target fluxes were correct, the bloth image fluxes came from the target image beam sizes. Moved nowhere. 13105. March 10, 2010 UVLOD, FITLD Eric Changed UVLOD and the UVLOD portion of FITLD to check each vis record for pure 0.0 in all spectral channels of a polarization/IF and to flag such groups. They report the count at the end. Changed MAXSOU in FITLD to 2500 from 500. Moved nowhere. 13106. March 10, 2010 BPLOT Eric Changed it to check a BP table record for valid data before including it. Otherwise, it plotted fully flagged antennas which is a waste of time and space. Moved nowhere. 13107. March 11, 2010 MNJ Eric When a new AIPS_ROOT file is added to the automatic move to $AIPS_ROOT and then define with AIPSROOT.DEFINE, it must be first in the list. Otherwise others may be encountered first and the new one will not be present. This causes the define for those encountered first to fail. Rearrange UPDUPDATE. Moved nowhere. 13108. March 11, 2010 EVAUV Eric Tried to open LUN1 (= 23) twice in the DFT routine - may well not be tested. Corrected code anyway. Moved nowhere. 13109. March 12, 2010 ACCOR Eric Corrected the Fortran to use a big enough summing array. It should be 2 * MAXANT * MAXIF or 16200. It was only 2000. Changed the help file and explain to discuss its use in more detail: "Some correlators apply this correction to the data before writing them out - notably the EVN JIVE correlator and correlators used for the Australian LBA. The VLBA hardware and software (DiFX) correlators do NOT apply this correction to the data. Therefore, ACCOR is required for the VLBA correlators and any others that do not apply the correction. ACCOR should be benign (do nothing) for those correlators that do apply the correction prior to reading the data into AIPS. Note that you can always run ACCOR and look at the SN table produced with PRTAB or SNPLT to see if it was benign or not." Moved nowhere. 13110. March 14, 2010 FITLD Eric Changed POPSDAT and NEWPARMS to add adverb ANTNAME*8(50). Addded this adverb to FITLD and revised code to pick up the adverb and set an initial list of antenna names (and therefore numbers) for when a new IDI data set is read. This will allow the user some control over antenna numbering but only for FITS IDI data sets. New help file for ANTNAME also. Moved nowhere. 13111. March 16, 2010 self defense in modeling Eric Changed ALGSTB to check for UMATS zero and to fill in the UMAT found for the current field with messages. Changed UVPROT to drop an odd line that set the max in U to the max rotated U. Added line to set the max rotated U to the max input U if the rotation appears to make a short maximum (less than 0.9 max input u). These alone will solve UVCON. Moved UVPROT to patch March 17. 13112. March 16, 2009 SETGDS Eric Changed it to ignore the warning from CATIO that the UV file is currently being written when the header is read. UVCON needs this since the UV file is the output not input file. Moved nowhere. 13113. March 17, 2010 UVCON Leonia As it was discovered recently, UVCON has not worked with image as a model at TST (DEC10) AIPS version. Eric has fixed the problem (see the previous fix 13111). Additionally the two first inputs at the subroutine SETGS are changed now from zeros to more general parameters. Moved nowhere. 13114. March 17, 2010 SPLAT, SPLIT again Eric The alternate reference pixel was computed ignoring the fact that the reference pixel was corrected for BCHAN by UVGET. Moved to patch this date. 13115. March 18, 2010 FITLD Eric Corrected the HI file writing of ANTNAME, making it smarter so that as many strings as possible appear on each line. Corrected the ANTNAME initial handling to blank fill the character array and to drop duplicates where possible. The names that appear in the HI will not be VLBA but the expansion of that string. Moved nowhere. 13116. March 19, 2010 ANBPL Eric Added adverb FACTOR to control the plot symbol size. Corrected the top label which used uninitialized pointers in some modes. Corrected the reading of the antenna file to deal with missing antenna records (missing antennas). Moved nowhere. 13117. March 23, 2010 UVLOD, FITLD Eric Added per cent line to the count of records flagged for pure zero. Moved nowhere. 13118. March 23, 2010 REWAY Eric New task of interest primarily for EVLA. It attempts to determine the data weights from the rms of the spectral channels, using ICHANSEL to avoid edge effects, RFI, and real line signals. For the weights to be calibrated, the data being fed REWAY must be calibrated or the calibration applied while running REWAY. Note that REWAY can be a substitute for a one source at a time SPLIT. The simple option determines the weight in each time, IF, and polarization product independently. Alternatively, the task can average the data in time and then solve for the weights on an antenna basis, still treating each IF and polarization product independently. These antenna weights may be smoothed in time with a Gaussian before being applied to the data. Moved nowhere. 13119. March 24, 2010 FITLD, UVLOD Eric Changed counting of all-zero spectra to count all spectra that are fully flagged on input and to examine only unflagged samples for pure zero. This the flagged for pure zero count will not report spectra that were already fully flagged but may include spactra (should they occur) that are partly flagged with all unflagged samples pure zero. Moved nowhere. 13120. March 25, 2010 CALIB Eric Changed the data averaging code to make it have 3 modes: full vector, vector over channels plus scalar over time, and full scalar. These are CPARM(5) 0, 1, 2 resp. The mixed mode is suitable for channel-independent atmospheric phases causing things to jump up and down. A side benefit of this change is that the averaging code is now much cleaner and easier to read. Moved nowhere. 13121. March 26, 2010 FITLD, UVLOD Eric Continued the upgrading of these flagging counters - they now also separate cross correlations from auto correlations. Moved nowhere. 13122. March 31, 2010 FILLM, INDXR Eric Wrote new subroutine KBOPAC to implement Josh Marvil's method of determining opacities. It uses the old method of seasonal and surface weather to figure out a K Band opacity. Then a table translates that to opacity as a function of frequency. This is now applied on a per IF basis, especially useful for the wide range of frequencies in the EVLA. Changed FILLM and INDXR to use the new method by deafult. The old method remains available by setting BPARM(1) >= 100. Moved nowhere. 13123. March 31, 2010 DLESQR (DELZN) Amy The matrix size in DLESQR was too small such that if someone tried to run DELZN on with more than 10 antennas DELZN would die. Increased the matrix size from 50 to 500. DLESQR is just a double precision version of LEASQR and all the messages etc., still called it LEASQR, changed all those to DLESQR. COMLNKed DELZN, ATMCA, RMSD and PBEAM. Moved nowhere. 13124. April 5, 2010 CVEL Eric The logical variable ISVLBA got lost last September 20 and so VLBA shifts were based on an antenna position rather than the center of the Earth. The VLBA correlators rotate everything to 0,0,0 to make life easy on CVEL among other things. For other correlators we have to guess to which antenna the velocity was shifted inside the correlator. That guess is antenna 1 except for EVN for which it is antenna 2. NOte that any renumbering of antennas makes a mess of this. Moved to 31DEC09 this date. 13125. April 5, 2010 UVLOD, FITLD Eric Changed pure zero message to have nothing come out if there were no pure zero of any sort. Put number flagged in HI. Moved nowhere. 13126. April 6, 2009 VPLOT Eric Made code to plot scan breaks in the data. APARM(8) > 0. Moved nowhere. 13127. April 6, 2010 INDXR, FILLM Eric Added EVLA Memo 143 to help files. moved nowhere. 13128. April 7, 2010 CVEL Eric Changed CVLDOP to antenna number 0 to imply center of Earth not the Common variable ISVLBA. Changed CVEL to set up antenna number to CVLDOP based on array names - then consult user adverb APARM(10) to override if set to 1, 2, or 3. Moved nowhere. 13129. April 8, 2010 INDXR Eric Discovered that the AGAINS array in memory allowed only 8 bands whilst the lower read routines returned band numbers 1-10. This led to troubles for P band (using Q values) and Q band for antennas that did not have any values of their own. Also discovered that the band ID routine would never return KA band but used Q instead so KA had the same troubles as Q and used Q band values. Moved to 31DEC09 this date. 13130. April 8, 2010 INDXR Eric Changed the antenna gain application routines to use the generic antenna gain whenever an antenna-specific gain is not available for a particular antenna. Moved nowhere. 13131. April 9, 2010 ANFIND Eric Added new spelling for a bunch of antenna stations there is a rogue program somewhere named MK3IN that uses different station names than everyone else and that did not get the message about the handedness of AN files. Added names to ANCHECK.001 RUN file also. Also changed ANREFM so that it sets XYZHAND='????' if it recognizes nothing. Moved nowhere. 13132. April 9, 2010 MK3IN Eric I did not fix the handedness of the AN file. The comments make it clear that the internal system is left handed so fixed the writing of the AN file to reverse By. Changed XYZHAND from blank to RIGHT. Moved nowhere. 13133. April 10, 2010 CASA's WX tables Eric CASA messes up at least two column labels and the format of the time column. Put extra code in WXINI and TABWX to detect these failings and work around them. Moved nowhere. 13134. April 12, 2010 WETHR Eric This task does not use subroutine TABWX for good reasons. Made change in code to detect the wrong time format (signaled by WXINI) and work around it. Moved nowhere. 13135. April 12, 2010 BPLOT Eric Added some counters to avoid having the task pause when DOTV true and the last plot has appeared. In some cases, e.g. a limited list of ANTENNA, it did not know previously that it was done. Moved nowhere. 13136. April 12, 2010 PRTAN Eric Added option DOBTWEEN to ask for a second display of antenna locations. The locations displayed the second time are wrt the average of the antenna locations in the file or for the EVLA wrt the traditional array center used withe the VLA in FILLM. The option is ignored for old VLA data sets. Moved nowhere. 13137. April 13, 2010 POSSM Eric Made it smarter to lower NPLOTS before it starts if there are fewer per page than requested. Located a number of places to keep track of times better so that even with non-zero SOLINT the task is likely to know when it is completely done and so not ask about continuing to the next page. Moved nowhere. 13138. April 14, 2010 REWAY Eric Found three things to fix: (1) the weight application routines did not check whether the data were previously flagged and effectively unflagged them. (2) the antenna-based fitting routine needed to be fed magic blanks for baselines lacking valid data rather than 0 so the 0's were believed to be real. (3) the aparm values were not written to the HI file. Moved nowhere. 13139. April 20, 2010 DOUVCOMP Eric Changed POPSDAT.HLP to make DOUVCOMP start -1. Edited the help files for DOUVCOMP, UVLOD, FITLD to add more negative information about DOUVCOMP. Moved nowhere. 13140. April 21, 2010 BPASS Eric The call sequence to FRQUPD was not corrected last year in this task. Has no VLB person run BPASS? Moved to 31DEC09 this date. 13141. April 21, 2010 VPLOT Eric Found bug in which the pointer for the first sample gets overwritten and so the first sample is plotted somewhere else if at all. Upped all the buffer parts by 10 time samples and now it does not happen. Fixed units for weights. Moved nowhere. 13142. April 21, 2010 UVFLG Amy Fixed flagging for shadowing. The hour angle was being misscalulated because the station longetude was being used instead of the array longitude. Also added a 10 second the end time used for the shadowing. Moved nowhere. 13143. April 22, 2010 FRING Eric This real*4 time business is a pain. Changed handling of scan start and stop to round outwards. There was some finite chance that records would occur 1.1 millisec past the apparent end of scan and so make a second solution of poor S/N. Moved nowhere. 13144. April 22, 2010 REWAY Eric Added option to smooth baseline-based weights. Add control over type of smoothing and to ignore pts > n sigma in smoothing. Added flagging of weights < or > set limits. Moved nowhere. 13145. April 23, 2010 REWAY Eric Correct help file - things are done on RMSes not weights and the conversion to weights only occurs at the end. Moved nowhere. 13146. April 23, 2010 VLA.GAINS Brian Butler Updated the VLA antenna gain curves with new values for X, K, KA, and Q bands. Fitting was done by Rick Perley. Moved nowhere. 13147. April 23, 2010 EVAUV Eric Changed Fortran and help to get 1. Added RPARM(30) output results array 2 Drop CMODEL - means sub-images not supported, CC preferred 3. Added list of plots in help 4. Improved error message strings 5. DFT modeling required a second BUFPUT to put the data back in the pseudo AP - it was dividing into the subtracted data rather than the original data 6. Find max-min of subtracted and divided over n sigma with n=8 for subtracted and 24 for divided. Was n=3 for both. 7. If value exactly matched the peak, changed hitogram bin number by 1 so that not counted as overflow. 8. Call MADDEX only when DOTV false - it was creating phantom PL files in the header 9. Plots of subtracted vis and gains made square or symmetric if they are naturally nearly so and extend outward a little. 10. Plot gain - 1 rather than gain. 11. Advance pointer through I/O buffer. It was replotting the same point many times and then skipping much of the data when reading subtracted and divided data sets for plotting. 12. Move print of #bad, average bad, etc to before GFINIS 13. Branch out on user request from plotting on TV. It would not stop otherwise despite what the prompt said. 14. Plot values of mean and rms of image on the image histogram plots. 15. Add solution for peak of histogram in plot routine so that the plotted fit would look right - i.e. match the histogram. Previously it plotted a Gaussian of correct width but with too large scale so it looked like it did not fit. 16. Fix axis labeling decision code to allow a max of 8 ticks rather than 9 which matches the assumptions elsewhere. The Y axis label could fall off the left sometimes. Moved nowhere. 13148. April 26, 2010 CALCODE Eric Added the EVLA translation table to CALCODE.HLP. Moved nowhere. 13149. April 26, 2010 INDXR Eric Added counters to determine if the data times are in the same range as the WX table times. For when the WX table is messed up by the EVLA or casa. Moved nowhere. 13150. April 27, 2010 UVLOD, FITLD Eric Made sure that apparent RAs are positive. Moved nowhere. 13151. April 29, 2010 REWAY Eric Added a new algorithm which seems preferable to the antenna-based weights one. If APARM(1) > 0, then find weights using a rolling buffer containing APARM(1) times. All times in the buffer are used in the robust rms fit and the resulting rms is assigned to the records/time of the middle of the buffer. The first and last APARM(1)/2 records also get the value assigned to the middle at that point so that the same number of samples go into each average. Change of source name or excessive time difference (APARM(8)) determines scan boundaries. The antenna-based algorithm is now run as APARM(1) < 0 and is thought to be less interesting. Moved nowhere. 13152. April 30, 2010 SNPLT Eric When plotting various X axis types like elevation, it was possible for the X value to be magic blanked. The task did not test for this and so had axis values of 3140.... Moved nowhere. 13153. May 1, 2010 REWAY Eric Added a second place to count times in the new method. The dynamic arrays were way too small for the rms smoothing and flagging and so the attempt to free the dynamic memory aborted the task. Added a counter and better test - data already flagged was counted as flagged by the smoothing and clipping. Moved nowhere. 13154. May 2, 2010 VPLOT Eric Discovered that its idea of the total number of plots assumed that all channel and IF plots were present. It did not attempt to plot the missing but it did not know that it was at the end unless the end happened to fill a page. Made a much more complicated count of total number of plots - matching the decision about whether to plot or not. Moved nowhere. 13155. May 2, 2010 REWAY Eric VERY GOOD TASK for finding questionable data. Fixed the smoothing ocde to not include samples exceedng the allowed weight range while smoothing. This makes it also exclude already flagged data (weight 0) also. Added progress messages every 100000 vis records - 50000 for the floating buffer method. Moved nowhere. 13156. May 4, 2010 REWAY Eric Rearranged the APARM(i) sequence. Changed default sigma cutoff to 4, added a cutoff based on the overall average and the sigma from there with difeault very large. Added the option to write an FG table on the input file based on the cutoffs. Put the final statistics about flagged records and Real/Imag > 1.5 in the HI file along with all adverbs. Made the antenna-based method always smooth by at least 1 second so that cutoffs may be applied. Moved nowhere. 13157. May 5, 2010 FRING Amy FRING would die of a segmention fault when clean component models where used. This was because FRING had an internal subroutine (CHNCOP) with the same name as an external subroutine which was called by UVMDIV with a call sequence that did not work for the internal version of CHNCOP. Changed the name of FRING's internal CHNCOP to CHNCPY. Moved nowhere. 13158. May 5, 2010 LGEOM, HGEOM, PGEOM Eric Changed these old chstnuts to use dynamic memory to put the entire input image plane in RAM. They still lose pixels to edge and blanked pixel expansion from the Everett interpolation, but they no longer lose a lot of the image to memory limitations. Moved nowhere. 13159. May 5, 2010 TVWIND Eric Changed the roam window setting routine to try to maximize usage. It was doing 3x4 frames for a 4096x4096 image when 4x4 was available. Corrected the ROMODE format described in help files for TVROAM and ROMODE. Moved nowhere. 13160. May 5, 2010 VPLOT and friends Eric Changed VPLOT, ANBPL, CLPLT, CAPLT, and UVPLT to have adverb XYRATIO with default to fill the TV on DOTV true and default 1.0 for plot files. Changed INPUTS section of VPLOT and ANBPL to list the BPARM(1) and (2) options as is done with UVPLT. ANBPL had residual code which messed up defaults and did some scaling that was incorrect (was for u,v,w). Moved nowhere. 13161. May 5, 2010 REWAY Eric Changed help and fortran: if OUTFGVER <= 0, write no flag table. If OUTFGVER ends up = FLAGVER, write message and change OUTFGVER to make a new file in order to avoid a serious gotcha (the data are read twice and must use the same FG table whilst reading). Moved nowhere. 13162. May 6, 2010 TV SERVERS Eric Shortened the title used for XAS by UNIXSERVERS and XASERVERS so that one can see the essential characters in icon displays. Moved nowhere. 13163. May 6, 2010 POSSM Eric Improved the label of LOG10 plots to include the BPARM(5) scale. Fixed inputs section of help to admit to BPARM(5). Changed help description - which was already there - to add a suggestion that it be set to an integer. Moved nowhere. 13164. May 6, 2010 VPLOT Eric Add plot type 22 = log(amplitude) and allow -22 to do log(amplitude) and phase. Moved nowhere. 13165. May 6, 2010 REWAY Eric Clarified the WT table as an extension file in the help. Moved nowhere. 13166. May 6, 2010 TVFLG, SPFLG Eric Changed GRIDTB formats that report number of points included - I8 is no longer adequate! Changed to keep track of max source name length and use that in interactive displays. Fixed formats in SPFLG's gridding routine also. Moved nowhere. 13167. May 10, 2010 SAD, JMFIT, MFPRT Eric (1) Clarified the explanation of the mark characters in the help and added outside primary beam mark information. In Fortran, changed ( ) surrounding the peak flux uncertainty to ! ! when the object is outside the reliable part of the primary beam. (2) Changed PMFC.INC, MFINI.FOR to allow for 2 more columns to hold the primary beam correction factor and the bandwidth smearing correction factor. Changed SAD, IMFIT, JMFIT to fill in these factors if computed in the task. (3) Changed MFPRT adding an adverb to scale the fluxes by these factors when they are present in the MF Table and fixing the task to allow the 41 columns plus column 42 for velocity. Mostly the # columns was a parameter but 2 places at least used a hard coded 40. Moved nowhere. 13168. May 10, 2010 MCUBE Eric Clarify the wording in the help file a little. Moved nowhere. 13169. May 11, 2010 RNXUPD Eric The automatic NX table writer did not recognize gaps of 1 second as gaps and so did not do the right thing for EVLA data. Changed it to slightly less than 0.2 sec for gap recognition. Moved nowhere. 13170. May 11, 2010 LISTR Eric Changed the LIST function to forgive and deal with the absence of NX and SU tables. A SCAN listing of single-source files is also desirable. Moved nowhere. 13171. May 13, 2010 FITTP, FITAB Eric Upgraded the data set not found message to show catalog number range (when there is one) and be less opaque as to its meaning. Moved nowhere. 13172. May 19, 2010 UVGLU, ALLTAB Eric Found that UVGLU was copying the NX table from the input file to the ouput even though it would not necessarily be correct. Corrected UVGLU to comput NX tables on the fly and changed subroutine ALLTAB to never copy an NX table. Moved nowhere. 13173. May 20, 2010 VLANT Eric Corrected code to handle antenna station names like N04 in the baselines.evlais files as well as in the antenna files. Put checks in for date loops - VLA ends with 2010, EVLA starts with 2010. Put VLA.ant.2010 in $AIPSTARS since it is complete, we are now working on EVLA files. Moved nowhere. 13174. May 20, 2010 BPASS Eric Changed some of the confusing meaning of BPASSPRM(5). Now +1 means no channel 0, 0 means record-by-record channel 0 in amp and phase, -1 means record-by-record but only in phase, and -2 means amp and phase using the time averaged visibilities. Moved nowhere. 13175. May 21, 2010 ANBPL Eric The scaling was not correct for weights since it started with PMIN=1000 which may be too low for the min weight. Changed to large numbers to avoid this issue. Also squeezed plot closer to top of each panel. Moved nowhere. 13176. May 21, 2010 QEDIUTIL Eric The max/min routine had some funny fixup code when the range was small. Changed it to allow rather small ranges. Moved nowhere. 13177. May 24, 2010 FARS Leonia Add convolution of the found clean components. The convolution is carried out under control of: APARM(7): convolve or not convolve; Aparm(8): convolve using Re of RMTF or Gaussian as the convolution function; Aparm(9): The full width of the used Gaussian. Moved nowhere. 13178. May 25, 2010 REAMP Eric Changed it to have CPARM(i) scale subarray i - so it is no longer a pure FUDGE clone. Moved nowhere. 13179. May 26, 2010 ANREFM Eric Changed it to be tolerant - but with a warning - about AN tables missing expected columns. Lots of link edits. Moved nowhere. 13180. May 26, 2010 RLDLY Eric NEW TASK - solves for the delay difference between the R and L polarizations. This is needed to fix slopes in the phase in RL and LR. This will proplace the great kludge known as VLBACPOL eventually - after wider testing. The REFANT must be specified but it may be any antenna, not just the highest or lowest in the data. Moved nowhere. 13181. May 27, 2010 FARS Leonia Small correction of the May 24 change Moved nowhere. 13182. May 29, 2010 NX table writing Eric This is tricky to get right. Changed DRNX.INC to include RNXEPS the min time interval to count as a gap between records in the same scan in days. Also renamed NUMANT to RNXANT to avoid conflicts with other includes. Changed RNXINI to set RNXEPS AND RNXANT and RNXUPD to use RNXEPS rather than the 0.2 sec TEPS it had before. Changed RNXUPD and RNXCLS to use RNXANT rather than NUMANT. This allows a calling task to mess with RNXEPS between RNXINI and the first RNXUPD call. Changed UVAVG to do just this since it knows the normal record interval. Changed RNXUPD to require 2 gaps before deciding for sure what the real scan break gap might be. Moved nowhere. 13183. May 29, 2010 INDXR Eric Changed it to use the code from RNXUPD to find the normal and scan break gaps in the data when CPARM(10 = 0. Moved nowhere. 13184. May 30, 2010 PLOTR Eric Added option of fitting orthogonal polynomials to the sets of data being drawn. Up to 8th order is allowed but it may not be possible to find orthogonal polynomials for such high orders. Moved nowhere. 13185. May 31, 2010 RSPEC Eric NEW TASK: plots spectrum of rms noise as a function of channel much like ISPEC but the noise rather than the signal. It has a median statistic approach and the more normal robust rms approach with a conservative and an agressive form. Moved nowhere. 13186. June 1, 2010 EXTLIST, PLGET Eric Made a code number for RSPEC (34). Added it to the list in AU2A for PLGET and corrected the number of adverbs for a number of tasks. Changed AU8A (verb EXTLIST) to recognize RSPEC and to have more adverbs for PLOTR. Moved nowhere. 13187. June 2, 2010 QEDIUTIL Eric I mistook the units in some of the data - it really is 1.0e-9 sec rather than 1 ns and some strange very small unit for rates. Changed the min-max routine again to take data type into account. The previous problem was an inability to blow up a small range and that should now be fixed for delays and rates also. Moved nowhere. 13188. June 8, 2010 FARS Leonia Add residuals (having subtracted the complex clean components) to send to the output. Help file is updated. Moved nowhere. 13189. June 21, 2010 UVLOD, FITLD Eric The computation of apparent coordinates from epoch coordinates was failing in the case of FITAB-written files because the SU table occurred before the observation date was known. Moved the call to the end at which time it is known. Moved nowhere. 13190. June 21, 2010 freq smoothing with cal Eric To save time, the low level routines DATCAL, DATBND, and DATPOL only calibrated the channels BCHAN to ECHAN. They must calibrate this range plus a smoothing support size in both directions. Corrected them to do this. Why has this not been noticed in 20 years???? Changed UVGET to init these as BCHAN and ECHAN while SETSM will alter them if smoothing is being done. Moved nowhere. 13191. June 22, 2010 FARS Leonia Correct the small error near the precise definition of the amplitude maximum. Moved nowhere. 13192. June 23, 2010 CookBook Eric Put back the draft of Appendix E with the Table of Contents and Indexes updated. Did html and COOKBOOK.PDF also. Moved nowhere. 13193. June 28, 2010 Image Scaling and Clean beams Eric In a data cube, the planes are made each with its own Clean beam to obtain best resolution at least in many cases. So we need to (a) keep track of all the Clean beams used and (b) change the scale of all the planes to be in the units of Jy per Clean beam in the header rather than the natural Clean beam. Thus CGINI New Table for Clean Beams: init TABCG New Table for Clean Beams: I/O TABLEUTIL New OCGINI and OTABCG for OOP access. IMAGR Add HDRBEAM keyword into Clean facet objects QCLEAN Add CLSCAL and CLCGEX subroutines to un-scale or scale each plane by the ratio of the actual Clean beam to the header clean beam. The latter also adds a line to the CG file at each plane. Local common holds HBMAJ et al, init on first CLNINI call. Set on first output image plane or when product type changes. WRPLAN Add a SCALE to the call sequence and scale the output image plane. MCUBE Pick up the input Clean beam areas and use ratio to scale output plane(s). Add subroutine to write one row of an output CG table for each plane written to the output cube. FQUBE Corrected oversight - make sure that the output has a FREQ axis - FQUBE requires it on a second execution. Like MCUBE for scaling and writing CG table. Moved nowhere. 13194. June 28, 2010 TVFLG, SPFLG Eric Write number of FG records written to history file even if details are not written. Moved nowhere. 13195. June 28, 2010 CONVL Eric Changed Fortran and comments in the help file. It now reads the CG file if present (and FQ file to get freq to get CG entry) and uses that Clean beam to determine the Gaussian to use in convolving the image. The result is a cube that is truely all at the same resolution unlike what we had before. What we had before were planes at different resolutions both before and after CONVL or image residuals at different resolutions and intensity scales when the Clean components were restored with a forced BMAJ. Moved nowhere. 13196. June 29, 2010 AIPSLetter Eric Put back the June 30 AIPSLetter Moved nowhere. 13197. June 29, 2010 CONVL, MCUBE, FQUBE Eric CONVL was limited in the ability to find CG entries by the accuracy of frequency computation. Changed it in CONVL, MCUBE, and FQUBE to try to be more accurate. Also made the search of the CG table for an entry have a second try with a more lenient error allowance. Moved nowhere. 13198. June 30, 2010 UVUTIL Eric Changed U2IDES so that output images will have frequency reference pixel 1.0. This makes the header frequency and the header Clean beam be for the same frequency. Moved nowhere. 13199. July 1, 2010 UVFLG Eric Added option APARM(7) to control the number of INFILE flags printed and fixed up that printing to include antennas, IFs, and times as well as the reasons. Improved wording in help file in other options. Moved nowhere. 13200. July 2, 2010 medians Eric Bryan Butler found a selection algorithm by Niklaus Wirth in his book "Algorithms + Data Structures = Programs" p 84 that finds the Kth ordered element on an array. I have written it into a file called MEDIAN with a function MEDIAN that calls a function KWIRTH with appropriate arguments. It turns out that this is very much fater than a full sort with RSORT for this function. Changed to use the new MEDIAN: RSPEC Use the new MEDIAN. Also make BLC(1,2) to TRC(1,2) a region of exclusion in finding the rms. BLC(3-7) to TRC(3-7) is still an inclusion region. Changed wording in the help file. REWAY Added OPTYPE MEDI and MEDR to do simple and robust method with MEDIAN as the statistics finder. Changed counters to separate Real/Imag > 1.5 from Imag/REal > 1.5. UVPHAS Use the new RSORT. One of the two uses took the 3/4 point in a descending array and called it the median coherence - it now does the correct median (1/2 point). Used by task PHASE. FLAGR Changed MEDIAN function to MEDIUN and have it call new MEDIAN. FINDR Changed MEDIUN to be MEDIAN and dropped local copy. Changed local MEDIAN to MEDIUM and have it call MEDIAN rather than RSORT. CLCAL The routine that applies an SN table to a CL table did a number of medians which are changed to use the new method. DTCHK and MWFLT were not changed. Both use several points in the sorted array so a sort is probably still the way to go. DTCHK is not used and MWFLT has a wide range of options that use various aspects of the sorted array. Moved nowhere. 13201. July 5, 2010 UVHGM Eric For some reason, the fancy fitting routines sometimes think that the arrays are orthogonal. Added a simplistic mean and rms fit to use in such cases. Found the reason for the failure on Intel 64-bit compilations: an array was improperly declared (3,3) when it should have been (14,3) given the other probably excessive parameters in the task. Moved nowhere. 13202. July 6, 2010 DOBAND, UVPRT Eric Changed BPGET to be more forgiving when interpolating bandpasses. If one of the 2 bandpasses is good in at least some channels and the other is fully blanked, then the good one will be used much of the time. Changed UVPRT to assume scaling based on adverb DPARM much like PRTUV rather than finding it on its own. DPARM(1) < 0 still orders it to find the scaling. It was taking too long to look at EVLA data. Moved nowhere. 13203. July 6, 2010 SPLAT Eric Discovered that the local version of RNXUPD had not been upgraded when it was changed May 29. In particular NUMANT(subarr) has become RNXANT(subarr) and so the compiler got upset. Moved nowhere. 13204. July 7, 2010 LWPLA, PLCOLORS Eric Mention procedure TVCOLORS apropos PLCOLORS in the two help files. Moved nowhere. 13205. July 7, 2010 Plot tasks Eric Changed: AU2A Changed number of adverbs for UVPLT, VPLOT, ANBPL, CLPLT, and added CAPLT to lists (verb PLGET). GINIT Added RSPEC and CAPLT to comments AU8A Verb EXTLIST - corrected parameter addressing for UVPLT and VPLOT. Corrected number parameters also for CLPLT and ANBPL. Changed axis type lists to have separate lists for VPLOT, UVPLT, and CLPLT. Added CAPLT to CLPLT with minor differences on IFs. CAPLT Changed plot type code from 22 (CLPLT) to 35. CLPLT Changed Number parms to 306 (305 was wrong). Moved nowhere. 13206. July 8, 2010 ANREFM troubles Eric DBCON with DOARRAY = 1 could mess up the input headers when the AN file needed reformating. Changed: LUNTMP New function routine to return an available LUN. RMEXT Use LUNTMP not LUN=28. ANFIND Add 'MWA' as a known array ANREFM Do not use the provided CATBLK since it may have been revised on us. Get a safe one from CATIO for creating corrected AN files. Moved to 31DEC09 July 9. 13207. July 8, 2010 copy BP tables Eric Changed BPFSEL and local copy of it in UVDEC. They were deleting partly flagged (one polarization, not other) BP table records. In fact one must keep even fully flagged solutions or the application routines will behave differently after a copy. Moved to 31DEC09 July 9. 13208. July 9, 2010 REAMP Eric Changed adverbs so that PBSIZE gives up to 64 scaling factors. OPTYPE selects whether these are by subarray or by IF. If they are by subarray, you may select one IF to apply these numbers or all IFs. If they are by IF, ten you may select one subarray or all to apply the factors. All IFs and subarrays are copied in any case. Moved nowhere. 13209. July 12, 2010 CALIB Eric Added more gain normalization options. If CPARM(2) = 2, the normalization is different for different subarrays. CPARM(2) = 3 means also separate normaliztion by IF. And CPARM(2) = 4 means separate R and L additionally. Moved nowhere. 13210. July 13, 2010 Spectral indices Eric Spectral index by itself does not require careful definition, but when curvature terms are added this changes. We now define the reference frequency as 1 GHz and R = log10 (freq/1.e9) log10 (T/T_0) = a R + b R^2 + c R^3 + ... where a is the usual spectral index and b,c,... are the curvature terms. T_0 is T at 1 GHz. This requires changes: SPIXR Changed to fit this formula rather than base e with some reference frequency as the reference PBUTIL Changed to interpret the spectral index and curvature image by this formula. UVMOD Changed to use this formula for spectral indices but still to interpret the component fluxes as being those at the reference frequency. Changed help files for SPIXR, UVMOD, IMAGR, and OOSUB to point out this units change. Moved nowhere. 13211. July 13, 2010 BPASS spectral indices Eric Using the above definition, created new adverb SPECURVE which is dimensioned 3 but you will probably only use the first value (SETJY only has the equivalent of SPECURVE(1) for the models of 3C286 and friends). Added this adverb to BPASS and made it apply base 10 referenced to 1 GHz spectral index and curvature to the bandpass solutions. Note that they may be normalized after that under control of BPASSP(10). Changed help files for BPASS, POPSDAT, and new SPECURVE and run file NEWPARMS. Moved nowhere. 13212. July 13, 2010 UVRFI Leonia UVRFI has had an ability to clean RFI components using only one or two components with GAIN=1. Now inputs parameters: GAIN and FLUX are added to allow making the CLEAN with any values for these parameters. It may improve the RFI mitigation at some cases. Moved nowhere. 13213. July 14, 2010 CLIP, FLGIT Eric CLIP: Examined the APARM(7) option and changed it to extend APARM(7) channels from the offending channel. It does this even if that takes it outside BCHAN-ECHAN. The flag all channels if more than APARM(9) are flagged really does apply to all channels. CLIP: Help - corrected a reference to APARM(8) whcih should now be (9). Clarified the extent of the APARM(7) and APARM(9) options. FLGIT: Added BPARM adverb using BPARM(1) to extend a bad channel +- BPARM(1) channels and BPARM(2) to flag the full spectrum when >= BPARM(2) channels are flagged (icluding previously flagged). Moved nowhere. 13214. July 16, 2010 SOUSP Eric New task: fits spectral index parameters to the fluxes in an SU table or in the adverbs. Plots in log-log and linear forms are produced optionally. The SPECINDX and SPECURVE adverbs are returned for use in BPASS. A new subroutine DFITPN was made from the single-precision POLYIN - it mostly was a simple change to double precision everywher but a weught array argument was added. Moved nowhere. 13215. July 19, 2010 install.pl Eric 1) The question after Screen 11 tried to go to SCREEN_12 which does not exist. Changed that to Screen 11 and allowed N and n as answers to go to it as well as B and b. 2) Changed SCREEN_11 to leave out advanced when binary install. Moved nowhere. 13216. July 20, 2010 BPASS Eric The spectral index scaling did not take into account that a spectral index was already applied at the reference frequency of each IF by the fluxes in the SU table. We need thus to scale by correction(i,IF) / correction(i0,IF). Moved nowhere. 13217. July 20, 2010 POSSM Eric Corrected yet another way in which it could exit when the user asked it to but without clearing the read status. Moved nowhere. 13218. July 21, 2010 BPASS Eric Changed it to insert the full spectral index curve for known sources - 3C286, 3C48, 3C147, and 3C138 AND 2C295. This may be prevented by setting SPECINDX to something other than 0 and 0.0 will be used iff SPECINDX < -5 or > +5. Moved nowhere. 13219. July 21, 2010 SETJY Eric Add the SPECINDX and SPECURVE adverbs. They are used in OPTYPE blank mode along with ZEROSP which becomes the flux at BIF with OPTYPE ' ' or at 1 GHz with new OPTYPE 'SPEC'. Moved nowhere. 13220. July 22, 2010 HELP CALCODE Lorant Added a few more that are now propagated through the system. Moved nowhere. 13221. July 26, 2010 BPASS Eric The spectral index correction for the reference channel was computed incorrectly (used channel N+1) so the normalization of the correction was off. Moved nowhere. 13222. July 27, 2010 PLOTR Eric Moved the orthogonal fitting option to DPARM(i) < 0 and added ordinary polynomials with DPARM(i) > 0. Moved nowhere. 13223. July 28, 2010 HGEOM Eric The dynamic memory change meant that a test on whether pointers came out right needed to be changed and it was not. This led to asking for dynamic memory 50 times and finally dying. Fixed the test. Moved nowhere. 13224. July 29, 2010 SPECTRAL Eric New scalar adverb added to POPSDAT and NEWPARMS. Help file now written yet - awaiting its use in PCAL. Moved nowhere. 13225. July 30, 2010 SPECPARM Eric More spectral index parameters in POPSDAT and NEWPARMS. Also putbck SPECURVE help file and improved SPECINDX help file. Fixed UVPRT help file so that it would not advertise options that do not exist (Real - imaginary and random parameter printing). Moved nowhere. 13226. August 5, 2010 TRUEP Eric A new task to determine the true antenna polarization from a data set in which one of the antenna's horn has been rotated by a known angle. This is mainly a task for commissioning telescopes but it got lost once and so needs to be in the system. Moved nowhere. 13227. August 6, 2010 UVMOD Eric The July change caused spheroiodal and uniform disk models to fail due to a change of variable type in some but not all places. Moved nowhere. 13228. August 9, 2010 AN FIND Eric Updated ANFIND to include WSRT as an array. Moved nowhere. 13229. August 12, 2010 UVLOD, FITLD Eric Added subroutine SUCHCK to change source names as needed so that source name is unique in the file. Moved nowhere. 13230. August 15, 2010 New tables Eric Created: PDINI Init new PD (polarization D term) spectral table similer to BP tables. TABPD Read/write PD tables CPINI Init new CP (calibrator polarization) spectral table (I,Q,U,V for all channels and IFs by source) TABCP Read/write CP tables CPSEL Copy CP tables selecting channels, IFs, FREQIDs. PDSEL Copy PD tables selecting channels, IFs, FREQIDs. BPSEL Copy BP tables selecting channels, IFs, FREQIDs - revised version of BPFSEL correcting file closing. COPTAB Add CPSEL, PDSEL, revise BPFSEL UVCOP Add CPSEL, PDSEL, revise BPFSEL SPLAT Add CPSEL, PDSEL, revise BPFSEL UVDEC Add CPFSEL, PDFSEL versions of CPSEL and PDSEL with channel increment. Moved nowhere. 13231. August 15, 2010 Spectral PCAL Eric Changed PCAL.HLP to 1. Add adverb ICHANSEL so that spectral channel averaging and selection may be controlled - previously by all channels. 2. Add DOMODEL to control whether models are used and whether the source Q and U are solved for or held fixed. 3. Add SPECPARM(4,20) to specify I,Q,U,V spectral indices for up to 20 calibration sources. The I spectral indices are used in spectral mode even if DOMODEL is false. The spectral indices of source 1 are also used when PMODEL is used even in continuum mode. 4. Add SPECTRAL to say (> 0) to do a channel-dependent antenna D term solution or to do a single solution averaging the ICHANSEL selected channels. 5. Added INTPARM(3) to smooth the UV data after all calibration is applied since S/N for a single channel is not likely to be adequate. ICHANSEL is honored by this smoothing. Changed PCAL.FOR to 1. Add all the above adverbs and code to honor them. 2. In spectral mode, dynamic memory is used to make large arrays to hold the solutions for the Ds and possibly the I/Q/U/V as functions of channel and IF. 3. The initial subroutine has a UVGET open and close to test the data and set a few parameters after which the source information and the frequencies (FQ table) may be found. 4. The INTPARM fuction is set up. 5. PCLSOU is also called early to get the source fluxes, names, etc from the SU table. It was called rather later and repetitively. 6. The main working routine now has to read a PD table if any to get the initial D terms in spectral mode or use those read from the AN table in continuum mode. 7. It now loops over channel just inside the loop over IF, including in each loop just enough channels to do INTPARM smoothing. 8. It also copies the source flux array to another variable adjusted for spectral index. 9. There are lots of differences in how the initial values are picked up before the fitting routines and saved afterwards. 10. The big change however is to honor DOMODEL and not fit for Q and U when using models. 11. The data are averaged with 2 routines now - PCLAVG for continuum and PCLSMO for line modes. 12. Also lots of minor changes of message levels, indentation, GO TOs etc. 13. Added code to force the reference R to 0.0 (rather than 5.e-8 which is what came out). Also code to interpolate across channels that were flagged or left out by ICHANSEL/INTPARM. 14. Add new adverbs to HIstory. 15. To allow option to read existing CP table to use as a model channel by channel for the calibrators. 16. Added QUAL and CALCODE adverbs. 17. Added code to fit spectral indices to the SU table fluxes when SPECPARM is 0.0. 18. Adde code to delete scratch files that were otherwise increasing by 2 with each channel/IF (an old bug but not so noticable with only 2 IFs and 1 channel). Moved nowhere. 13232. August 15, 2010 Plot new tables Eric POSSM Add options to plot PD and CP tables, the latter as Q and I or as Q/I, U/I. The plot range was implemented by changing the data to be the max plot value or the min plot value. This causes points that should have been omitted to be plotted as valid and causes lies to be printed if the output option is selected. Fixed the plot routines themselve to omit out of range data rather than lie. Has local version of SOUFIL which fills the source tables even if all sources are requested. BPLOT Add options to plot PD with antenna number on the Y axis and CP tables (as Q/I, U/I) with source number on the Y axis. Added XYRATIO with standard defaults. Added PIXRANGE(2) < PIXRANGE(1) option to self-scale each little plot individually. Moved nowhere. 13233. August 15, 2010 Apply PD table Eric Changed: DSEL.INC Removed the polarization calibration parameters DPDC.INC New include for old and new pol cal data. SELINI Add include DPDC and init pointers for it. UVGET Clear dynamic memory if DOPOL > 0 and PDVER>0. DOPOL.HLP Values 6,7,8 to be like 1,2,3 but forbid line solution POLSET Add PDPC, get PDVER for DOPOL < 6, drop unused section of code for non-linear solution modes. Make dynamic memory for PD table and read in DTERMS. Make dynamic memory for continuum or line POLCAL and fill it with AN or PD values in SOLTYP 1. DATPOL Changed call seq to xxxPOL routines. Pix up POLCAL from dynamic memory rather than a fixed memory. VLBPOL DOPOL = 1 or 6 for linear solution. Pick up DTERMS or the antenna values depending on PDVER. Moved to POLCAL by copy rather than equivalence. ORIPOL Pick up DTERMS or the antenna values depending on PDVER. Moved to POLCAL by copy rather than equivalence. LXYPOL Fix error - used 4x4 complex matrix as a 2x16 matrix without proper equivalence. Commented out a section of code that changed POLCAL which was then promptly over-written. Pick up DTERMS or AN values as needed, copy to POLCAL. Moved nowhere. 13234. August 18, 2010 TVFLG, SPFLG Eric Added option to turn on/off axis labels. The step wedge is labeled if it is on, replacing the X axis (antenna, channel) labels. The time axis is suitably non-linear. Changed Fortran files and help and HLPxxFLG.HLP files. Moved nowhere. 13235. August 24, 2010 PLOTR Eric Allow UV as well as MA types for the file to which plots may be attached. Changed adverb names to allow up to 30 line types. Changed internal buffers to go to 30 types and 100000 total samples. Changed format of "sample label" lines so that one may include only those desired, not all 30. Moved nowhere. 13236. August 24, 2010 PD file Eric The phase difference column must have either 0 or Nchan*Nif values in it depending on POLTYPE. Changed: DPDC.INC Dimension of POLPD is now MAXCIF PDSEL Dimension of PHDIFF is now MAXCIF - must select on channels as well as IFs. PDINI Make size of column 5 depend on POLTYP and curb error message when length 0 expected. TABPD Move copying of PHDIFF into 1st polarization big loop and make it conditional. PCAL Change PD to PHDIFF and set it properly, getting and saving in PD file as well as AN file. It was not saved in line case previously. POLSET Pick up PHDIFF only when appropriate to POLPD and get the right size ORIPOL Pick up PD and PRREF and PLREF from correct channel in line case. POSSM Dimension of PHDIFF is now MAXCIF BPLOT Dimension of PHDIFF is now MAXCIF UVDEC Select by channel and IF on PHDIFF in PDFSEL Moved nowhere. 13237. August 24, 2010 RLDIF Eric Added new adverbs DOAPPLY, POLANGLE to POPSDAT and NEWPARMS. Added new help file DOAPPLY and POLANGLE and corrected SPECTRAL help file. Changed Fortran to add a lot of things including a path through new routines for line work and continuum with more than 1 source. Added routines to correct the CL/AN/SU tables for continuum solutions and BP/PD/CP tables for line work. Writing a PLOTR-capable text file was added as well. The task now knows the expected angle for the 4 major sources and can be told the angles for all sources. Moved nowhere. 13238. August 25, 2010 Auto NX files Eric Changed the MAXGAP algorithm again to prevent it going under one minute in RNXUPD. Changed SPLAT local version of RNXUPD also. Moved nowhere. 13239. August 25, 2010 CALIB Eric Changed some of the remarks about SOLINT in the help file. Use of 1 sec SOLINT with data mostly averaged to e.g. 10 sec can lead to bad things with the data not at the same times. Moved nowhere. 13240. August 25, 2010 UVAVG Eric Added option to put out all baselines in an averaging interval with the same average time. The default is to put each out with the averaged time for each. Due to flagging, the resulting times will vary and are carefully sorted into time order. Moved nowhere. 13241. August 25, 2010 REWAY Eric Changed it to pay attention to the incoming weights when APARM(9) <= 0. It will use the weights in finding the rms except for the median methods. It finds the average weight entering the rms and scales the rms by multiplying by square root of the average weight (effectively dividing by the average expected rms on input). It is those numbers that are smoothed and used to make flags. Then the incoming data weight is multiplied by this change in rms. If APARM(9) > 0, the incoming weights are ignored in computing the rmses and ignored in setting data weights on output. Moved nowhere. 13242. August 26, 2010 SCMAP Eric When HDRBEAM was added, the self-cal tasks were ignored and so failed. Changed QCLEAN to suppress errors messages on the parameter named HDRRBEAM and default it properly. Moved nowhere. 13243. August 26, 2010 INQSTR Eric The subroutine which asks a user for a string limited that string to a sensible 72 characters. Our users are, I guess, not sensible. Changed INQSTR to have an internal string length of 256 and to check the string read and return an error if the string is longer than the limit requested. Changed REBYTE to allow directory names up to 192 characters. Changed QEDIUTIL (EDITA, EDITR, SNEDT, SCMAP, SCIMG), TVFLG, SPFLG, and IBLED to be understanding about excessive length strings and repeat the question. Moved nowhere. 13244. August 26, 2010 POSSM Eric Fixed typo in Stokes labeling and missing line in inputs description. Moved nowhere. 13245. August 27, 2010 POSSM Eric Changed it to recognize the end of plotting on BP, CP, and PD tables so that the last plot gets horizontal tick labels even when the page is not filled. Moved nowhere. 13246. August 27, 2010 NOIFS Eric NEW TASK: Converts channels + IFs into only channels. Requires the IFs and channels to have sensible frequencies for doing this operation and the data had better get calibrated in or before this task. It drops all cal tables except a new null CL table. It averages overlapped channels (weighted) and does the right thing with CP tables. Corrected reformat message in CPSEL although this task has local version of CPSEL and COPTAB. Moved nowhere. 13247. August 30, 2010 SAD Eric When the deconvolution found a size too small, it set the flux and position of the output CC record to zero, rather than the widths and position angle. Added the right subscript to the zeroing. Moved nowhere. 13248. August 30, 2010 UVAVG Eric Added two more averaging options. GRID will write out the averages on a strict grid from time 0 in the data set and all baselines from each grid interval will have the time of the center of that interval. GRIT will use the same intervals, but each baseline will get the actual averaged time appropriate to it. The time setting was moved ahead of the NX table updating since the times greatly affect how that routine works. GRID now applies to BT sorted data if specified, otherwise GRIT is used with BT data. Moved nowhere. 13249. August 30, 2010 ATLOD Eric Changed upper limit to tranlated file name to 256 characters. Changed to pass user's input name to TAPIO for file opening. The fully translated name could easily be > 48 characters. ZTPOPN declared its name variable as *48, should have been *(*) although that should not have mattered here. Moved nowhere. 13250. August 31, 2010 CLCOR Cormax Reynolds / Eric For OPCODE 'ANAX', telescopes with EW XY mounts require a slightly different geometry in their correction compared to altaz and Nasmyth mounts. Cormac reynolds has provided us with this corrected code. Corrected comment in help file about supported mounts on September 1. Moved nowhere. 13251. September 1, 2010 POSSM Eric Changed APARM(1) so the >= 0.0 yields vector averaging. The scalar averaging default is costing lots of confusion. To get scalar set APARM(1) < 0. Moved nowhere. 13252. September 1, 2010 REWAY Eric APARM(5) and APARM(6) can lead to flagging just like APARM(7) and APARM(8) but in a data-dependent way. This effectively can provide data dependent defaults for the latter two. Tried to improve help file wording. Moved nowhere. 13253. September 1, 2010 PCAL Eric An array effectively changed name while the old name was then used as a scalar. For some reason, numerous compilers allowed a loop copying the array by its old name. The latest Intel compiler on Macs caught it. Moved nowhere. 13254. September 2, 2010 FARS Leonia Add default for CELL. The defalt is ~PI/(4*(LAMBDA_max^2-LAMBDA_min^2)). The description of the input parameters is improved. Moved nowhere. 13255. September 3, 2010 CATLST Eric The listing routine for MCAT etc had a reasonable upper limit of 32000. That is now being hit so rased it by x30. Moved nowhere. 13256. September 3, 2010 ATLOD Eric Changed it to use JPRECS rather than PRECES since the latter claims to be for B1950 only. This is solely to get the apparent positions listed in the SU table. Moved nowhere. 13257. September 3, 2010 FLATN Eric The default INSEQ is 1 which differs from the usual default. Changed help file to state this clearly. Moved nowhere. 13258. September 7, 2010 VLARUN Lorant Changed BPASSPRM(5) from -1 to -2 to follow new definition. Moved nowhere. 13259. September 8, 2010 PDVER Eric 1. Created PDVER help file and added the new adverb to POPSDAT.HLP and NEWPARMS.001. 2. Moved PDVER from DPDC.INC to DSEL.INC where DOPOL already was. 3. Changed POLSET to examine header for PD versions and set PDVER appropriately, including to 0 when there is none or DOPOL >= 6. 4. Changed BPLOT to use INVERS rather than BPVER. 5. Changed POSSM.HLP to emphasize that BPVER is used when plotting BP, BD, PD, and CP tables as well as when calibrating UV data. Added PDVER adverb. 6. Changed to add PDVER adverb and to have it make a new file from PDVER when correcting for the RL differences. 7. Changed PCAL to read in the highest PD table initially and to copy that to the output replacing solutions as appropriate with the new ones. 8. Added PDVER to REWAY. Changed REWAY so that the output files are always uncompressed. 9. Added PDVER to CALEDIT known parameters 10. Added PDVER adverb to help and Fortran in: AVSPC CLIP CORER DECOR DIFRL FLGIT FLOPM FUDGE LISTR NOIFS PHSRF SHOUV SPLAT SPLIT SUFIX TI2HA TIORD TRUEP UV2TB UVAVG UVBAS UVFIT UVFND UVHIM UVHOL UVIMG UVLIN UVLSD UVLSF UVMLN UVMOD UVMTH UVPRM UVPRT UVRFI FINDR FLAGR MAPBM RFI UV2MS BLCAL BLCHN BPASS CALIB CPASS DEFLG FRING KRING RLDLY BLING EDITR IMAGR OMFIT SCIMG SCMAP ANBPL CAPLT CLPLT FRMAP FRPLT IBLED SPFLG TVFLG UVHGM UVPLT VPLOT WIPER Moved nowhere. 13260. September 8, 2010 TI2HA Eric Changed it to write a header keyword TIMETYPE='HourAngl' in its output and to check its input for this keyword and value. If it find it in the input, it issues a message and does a simple copy. Moved nowhere. 13261. September 10, 2010 obit in aips Eric Added BDFLIST and BDF2AIPS to $SYSUNIX and corresponding help files. These stand-alone scripts use Bill Cotton's obit package to list an EVLA ASDM contents and to translate them into an AIPS file on disk. These obit programs remain under active development so you will have to keep your version of obit up to date. Moved nowhere. 13262. September 11, 2010 image models Eric Changed MODFIT and GRDSUB to ask for more AP memory. It was not getting enough and so omitting all components from a facet! Moved nowhere. 13263. September 11, 2010 FRPLT Leonia The FRPLT has not worked at least with DOSHFT=.FALSE. I have fixed it. Moved nowhere. 13264. September 13, 2010 OGEOM Eric Changed code and help file to retain the reference pixel if the output image is the same size as the input image. Otherwise, the images are scaled about their "centers" defined as NX/2, NY/2 + 1. The old code divided by the scaling values when it should have multiplied when computing the new reference pixel. Moved nowhere. 13265. September 13, 2010 Y2K Eric Changed Y2KSAVE and POPSDAT to have MPDIFS(5,20) rather than MAPDIF(5,8) as in DDT. Changed DDTLOAD to use MPDIFS, to check for constant output difference image to avoid quitting when IMSTAT has no non-zero pixels, to save and check all 4 image files in HUGE Clean, and to report all check results. Computed new master images with Intel 64-bit load modules in largish allowed memory. They are stored in the AIPS ftp area FITS/31DEC10. Moved nowhere. 13266. September 13, 2010 LISTR Eric In the new code for 'SCAN' that covers for not having an SU table, the codes for OPTICAL/RADIO were switched. Moved nowhere. 13267. September 14, 2010 RLDIF Eric Found typo type error causing the updated PD table to have zero rows. Moved nowhere. 13268. September 15, 2010 UVRFI Leonia UVRFI has been modifyed a lot during last time. This modified version is put in the system. The UVRFI.HLP was changed also in acordance of modification of the codes and tring to do it more understandable. Moved nowhere. 13269. September 15, 2010 DOPOL Eric POLSET needs to do all channels surrounding BCHAN to ECHAN needed for the spectral smoothing. Changed BCHAN to BCHANS and ECHAN to ECHANS. UVGET needs to call SETSM before calling POLSET as well. The value MVIS(1) needs correcting after SETSM to specify the channels including those needed for smoothing rather than just BCHAN to ECHAN. MVIS(1) does not seem to be used for anything much anyway. Moved nowhere. 13270. September 16, 2010 PD, CP Eric The automatic copying was not turned off on these types in SPLAT, UVDEC, AVSPC, and UJOIN which do their own table copying. (Subroutine COPTAB does this right for most cal applying tasks.) Put messages in UJOIN and AVSPC about discarding BP, CP, and PD tables. Moved nowhere. 13271. September 16, 2010 WIPER Eric Inputs did not get PDVER while the Fortran did - messing it all up. Moved nowhere. 13272. September 16, 2010 ANFIND Eric The code asked is ANAME(:3) = 'WSRT' which can never happen. So WSRT was still not recognized. It should be now. Moved nowhere - lots of link edits. 13273. September 17, 2010 QU Eric Changed FNDPOL, DGINIT, DGHEAD to allow for Stokes 'QU'. Changed POSSM to allow for STOKES = 'QU' and 'RLLR' and to allow for autocorrelation phase on many more Stokes values (those that include Q, U, RL, and or LR). Moved nowhere - lots of links. 13274. September 17, 2010 NOIFS Eric Fixed usage of header blocks - there are 2 here unlike FUDGE - one for the multi-IF input and one for the 1 IF output. The header for the NX table and the final vis count was wrong so they were lost. If the final vis count was substantially less than the input, bad things happened. Moved nowhere. 13275. September 17, 2010 PCAL Eric If there was no data in a spectral channel, that IF was terminated in the solver. Improved the no data found message but the already closed in FTAB messages are too hard to get rid of. Moved nowhere. 13276. September 20, 2010 REBYTE Eric Compressed UV data format was forgotten in REBYTE. This caused the imaginary and real parts to be swapped. Added test for this format and a new routine to walk through the buffers to deal with it. Moved nowhere. 13277. September 20, 2010 SWAPR Eric New task: swaps real and imaginary to fix the above bug. Moved nowhere. 13278. September 20, 2010 RNXUPD Eric The NX table needs to be updated with all subarrays when a later time is being written in a particular subarray. This bug led to an index record being written at the end for times much earlier in the file. This leads to strange things as the data moves on to other tasks. Moved nowhere lots of links. 13279. September 20, 2010 BDF2AIPS Eric Added x.doSwPwr=False to turn off application of the switched power measurements which are not reliable or understood. Moved nowhere. 13280. September 20, 2010 SPLAT Eric Fixed the local copy of RNXUPD also. Moved nowhere. 13281. September 20, 2010 VBGLU Eric VBGLU ignored the reference channel in the frequency computation for the FQ table. Moved nowhere. 13282. September 21, 2010 SAD et al Eric Changed SAD, IMFIT, JMFIT in the special CfA format (DOCRT=-4) output for the reference coordinate. There was very strange code in the declination. It was not right for either positive or negative declinations. Moved nowhere. 13283. September 21, 2010 SPLIT, SPLAT Eric When selecting a source to write out, the task did not differentiate on CALCODE so that two sources with identical names and qualifiers would end up twice in 2 multi-source output files (unless CALCODE was specified exactly). Now these tasks can differentiate on CALCODE as well. Moved nowhere. 13284. September 21, 2010 UVLOD, FITLD Eric Changed SUCHCK to change source QUAL to avoid duplicate names and to test that CALCODEs match before concluding that the name is a duplicate. The old code renamed sources and had a possibility of a pathological happenstance keeping it from achieving its goal. The new code loops to stop this chance. Moved nowhere. 13285. September 22, 2010 DADEVS Eric Changed DADEVS.PL and DADEVS.SH to also read a file $HOME/.dadevs.always of data areas that are included regardless of host name. Those with a + in column 1 come first in the list, those with a - come after all disks that are "required" and those matching the current host name and before any matching hosti of da=host1,host2,... Moved nowhere. 13286. September 23, 2010 DADEVS Eric Changed $SYSMAN/aips.l, AIPS.HLP for the above change to the logic of DADEVS. Also prepared LaCook2 for this change (it will be putbck when we are ready to putbck other CookBook changes). Moved nowhere. 13287. September 24, 2010 SNPLT Eric Added the new SY table to the plot options - can plot Pon+Poff, Pon-Pof, PostGain, Pon, Poff, Tsys/Tcal from this new EVLA table. This table is available via obit but the gain column is not properly labeled yet. Moved nowhere. 13288. September 25, 2010 Copy image tables Eric A lot of tasks did not copy tables for images. Besides the CC tables we now have FQ and CG tables which are truely needed for coordinates and the like. Changed IMAGEUTIL Make new IMCALT to copy all tables not in DROPTABS keyword. Much like UVCALT in UVUTIL Changed to copy all tables: TAFFY XSMTH HGEOM IMFLT IMLIN LGEOM PADIM SMOTH XTRAN CCRES FLATN IMCLP OGEOM OHGEO BLANK SERCH XBASL Changed to copy all tables except CCs: AHIST HISEQ SCLIM SUMIM WTSUM XSUM CUBIT FARS IMMOD MATHS MEDI MWFLT NINER PGEOM RMSD SHADW PASTE APCLN SDCLN XGAUS Changed to copy all tables except CG: CONVL Moved nowhere. 13289. September 27, 2010 TVFLG Eric Changed GRIDTB to ignore missing data better and only complain if none were found by the end. Previously, if channels 1-8 were flagged then the grid routine would complain bitterly and the task would die even if later channels were present. Moved nowhere. 13290. September 27, 2010 ELINT Eric Improved some formats which had + signs followed by negative numbers. Also fixed the history cards which did not use the carefully correct displayed information but instead wrote the results as if they were wrt elevation and simple. Now the message line or lines go to the HI file too. Moved nowhere. 13291. September 27, 2010 TRANS Eric The old code could hit disk very badly taking 3 minutes to change XYV to VXY followed by 120 minutes to restore the image to XYV! Changed it to use vast memory and to read as many planes as possible into memory to reduce the I/O count and raise the I/O size in row swaps. The vast memory is limited by the maximum allowed AP size so it can be controlled by the user. Moved nowhere. 13292. September 30, 2010 TVHELIX Eric NEW VERB: TV OFM scheme due to Dave Green (MRAO) implements a helix in color space so that the perceived brighteness is monotonically increasing. The verb differs from Dave's version in the choice of which button or cursor position controls what. Changed AU6, POPSDAT, NEWPARMS.001 and added a new help file. Moved nowhere. 13293. October 1, 2010 FILLM, VLA.GAINS Eric Changed VLA.GAINS to make the end time for old VLA gains 20091231 instad of 2100xxxx. Some VLA gains were being used instead of the EVLA ones for antennas not in the last flux calibrator run. Changed FILLM to use the default antenna ("0") rather than 1.0 for the gains if there was no entry for a particular one in the desired date range. This fix was made in INDXR on April 8, 2010. Moved nowhere. 13294. October 3, 2010 CLEAN Eric CLNOPN and CLNCLS both set MSGSUP and neglected to reset it or was tardy in doing so. Moved nowhere. 13295. October 3, 2010 FRING Eric Put in messages about the integration time found in the data. It is getting some very short ones for some reason. Moved nowhere. 13296. October 4, 2010 FRING, RLDLY Eric Add check of output of AVERT - if integration time < 0.005 then assume units are days rather than the seconds they are supposed to be and correct the units. Moved nowhere. 13297. October 4, 2010 PRTUV, LPOPEN Eric LPOPEN set DOCRT<=72 to the window width - it was supposed to be <72 only. PRTUV chaned DPARM(6) print out to displat times to 1 millisecond accuracy. Moved nowhere. 13298. October 5, 2010 VLANT Eric Changed it to delete the /tmp/... file after reading it. Otherwise, one may not have permission to write over it later when a different users runs VLANT. Also found a place that did not set the name for evla. The thing did not fail as one would hope because there is a vla file too for this year. Moved nowhere. 13299. October 7, 2010 EQUINOX Eric Standard FITS has EQUINOX as a float but FITS-IDI and others call for 'B1950' and 'J2000' and variants in spelling. Changed: GETHDR to first try for a float. On error it gets a string and looks for 1950 and 2000 in the string. An error arises only when those fail too. SOINI Added comment that EQUINOX column may be character (SOINI incorrectly uses double) SOTAB Allowed both character and double before but used 'B1950.0' rather than '1950.0B' as required. Added that string as an option on read, changed write which never happens anyway. FPARSE VFUV.INC list: if numeric type fails and type is EQUINOX or EPOCH, then try string. Look in string for 1950 or 2000 and set X appropriately if found. Otherwise X stats DBLANK and an error arises. FITLD Global keyword parsing was wrong anyway so global equinox string never set properly. Changed to allow DP, SP, I formats or any string. Could have led to a wrong header equinox. Also added the special parsing (as in FPARSE) of EQUINOX to 3 separate subroutines from VFUV, VFIT includes. PRTTP Uses FPARSE plus a local TPARSE to manage tables. Foxed the latter to allow special parsing for EQUINOX and EPOCH. CALRD As FPARSE FIT2A As FPARSE GSCAT As FPARSE IMLOD As FPARSE SKYVE Added special code in the parsing of EQUINOX from the history information. UVLOD As FPARSE CONFI As FPARSE Moved nowhere. 13300. October 7, 2010 EDITA, SNEDT Eric Added SY tables to QEDIUTIL and tasks SNEDT and EDITA. This requires new run-time help files HLPEDISY and HLPEDISS. QEDIUTIL gets new subroutines EDITSY, EDGTSY (get data), EDG1SY (get a bit of data), EDIFSY (compute Pdif/-1 and Psum/-1), ans EDSMSY (smooth both Pdif and Psum). The smoothing array was made larger and a number of minor general issues were corrected. Moved nowhere. 13301. October 11, 2010 TYSMO Eric Added SY tables to the operation. Changed the adverbs so that APARM controls absolute value clipping, CPARM controls clipping on the difference between the datum and the median of its neighbors, and BPARM allows smoothing after all the clipping. For SY tables, the Psum, Pdif and Psum/Pdif/2 (Tsys/Tcal) are examined. The last of these cannot be smoothed since it is the ratio of the two measured quantities but cliiping on it is very productive. Finally required puting back $APLNOT/{SYINI,TABSY}, the two IO routines for SY tables. Moved nowhere. 13302. October 12, 2010 EDITA Eric Add to QEDIUTIL EDITSY subclass display and editing based on Psum/Pdif/2 = Tsys/Tcal. This is a very useful way to look at the SysPower data - both directly and as a relative difference with the running mean. Changed QEDIUTIL and help files for EDITA, SNEDT, HLPEDISY and HLPEDISS. Changed TVXPRT to handle blank menu items and fixed up the expert mode and a few other matters in QEDIUTIL. Moved nowhere. 13303. October 12, 2010 SLFIT Eric Updated the explain section of the help file to make the TK verbs nearly invisible and emphasize the TV approach. So far as I can tell, this is the only help file with such troubles. Moved nowhere. 13304. October 13, 2010 CCSEL Eric Raised the upper limit from 10000 to 100000 CCs. Moved nowhere. 13305. October 15, 2010 SPLIT, SPLAT Eric The EVLA seems to want to have the same source name with different CALCODEs or QUALs as different sources in the source table. This confuses SPLIT and SPLAT that use only the first 12 characters of the source name as the output file name. Changed them to test for duplicate source names and, in those cases, to put in the CALCODE and QUAL in characters 9-12. Help files have an added note about this. Moved nowhere. 13306. October 15, 2010 FGPLT Eric Rearranged the logic a good bit. Rather than store the flag info in enormous arrays (7 billion words would have been coorect!), the task now rereads the FG table for each page of the plot filling in whatever lines it finds. The initial read makes tables of the the all antenna, one antenna, and baseline flags for later use. If all flags in the time range are antenna based, the plot is done by separate code on that basis rather than on a baseline basis. Added BCHAN, ECHAN to limit flags to those that apply in that channel range. Moved nowhere. 13307. October 15, 2010 ATLOD Eric Bad format (F15.5) used for integer variable in history. Moved nowhere. 13308. October 15, 2010 TVFLG, SPFLG Eric Made labeling smarter when the wedge is also on. If the wedge is below, the wedge will be labeled and the image X axis labels will appear at the top. If there is no wedge they appear at the bottom. Things are reversed if the wedge is at the top (in the unlikely case that that happens). Moved nowhere. 13309. October 15, 2010 FRING Eric Added APARM(5) = 3 option to solve for 2 delays, one using IFs 1 through NIF/2 and the other using IFs NIF/2+1 through NIF. This will be useful for the EVLA where the delays differ between hardware IFs AC and BD but not between the more numerous spectral windows called IFs inside AIPS. Moved nowhere. 13310. October 17, 2010 SPLAT Eric My latest "fix" made a mess of the adverb values for source selection leading SPLAT to select the last source only when trying to assemble a multi-source output. Fixed to test source names only when not assembling which is the only time they need to be tested. Moved nowhere. 13311. October 17, 2010 FRING, RLDLY Eric Corrected 2 places in FRING where number of channels per IF was computed wrongly. Added DOIFS option to RLDLY to control whether a single solution, 2 solutions, or NIF solutions are done. Moved nowhere. 13312. October 18, 2010 SNEDT Eric Changed QEDIUTIL to apply the FG table to table data as they are read or reread. Changed SNEDT Fortran and help to add FLAGVER adverb: 0 -> do nothing. One must specify the desired FG table version. Moved nowhere. 13313. October 19, 2010 UVCOP etc Eric Changed GCINI to limit the number of polarizations to 2. Moved nowhere. 13314. October 20, 2010 FRING, RLDLY Eric the new option did not handle the data properly. In the new option, it has to copy the solutions from IF 2 to IFs NIF/2+1 through NIF and then from IF 1 to IF 2 through NIF/2. But the next routine which is called optionally in FRING should have moved the solutions from IF NIF/2+1 back to IF 2. This was not a problem when the only IFs option was to use all or one at a time. Moved to Socorro machines this morning. 13315. October 20, 2010 ATLOD Eric More errors with formats in HI writing - for parameters in DPARM that are used but not described in the help file. They seem to have to do with pulsar bins. Moved nowhere. 13316. October 20, 2010 SYSEL Eric Put back new routine SYSEL to select records from an SY table to be copied. Added calls to this in COPTAB and special COPTABs in UVCOP and SPLAT. Added SY to the lists not to be copied in the simple way. Moved nowhere. 13317. October 20, 2010 REWAY Eric Added support for CROSSPOW keyword - writes 0 if none or initially 1. Passes -1 along. CROSSPOW 1 is cross-power data and time integration only weights. CROSSPOW 0 has real weights and -1 has visibilities and real weights (from new TYAPL not yet released). Moved nowhere. 13318. October 20, 2010 BDF2AIPS Eric Changed BDF2AIPS and BDFLIST scripts to add the seconds onto the file name in /tmp. Users are taking < 1 minute between executions! Moved nowhere. 13319. October 21-22, 2010 Misc Eric 1. SPLAT had undefined variables in yesterday's putbck. 2. TABHDR parsed dimensions from any TDIM keyword but they are only for visibility data. Reworded error message and restricted parse to visibility data. (IMLOD, UVLOD, FITLD) 3. SNEDT: the application of flags to table data could return error code -1 if the last FG table record was flagged. This confused higher level tasks. (QEDIUTIL - also fixed formats) 4. FGPLT: the new code that gets the time range from the data fails miserably on TASAV files. Changed to get the time range from the flags in that case. Moved nowhere. 13320. October 22, 2010 NX table Eric Changed RNXINI to issue a warning about a pre-existing NX table rather than a quit writing level error. Changed UVFIX to write an NX table on the first pass through the data and skip it on any later passes. Moved nowhere. 13321. October 22, 2010 UVGET Eric Force DOWTCL to false when AN file says EVLA and header CROSSPOW >= +1. The latter now allows DOWTCL true for weights produced by TYAPL and REWAY. Moved nowhere. 13322. October 22, 2010 TYAPL Eric Put back CDINI and TABCD to support CD (CalDevice) tables. Added full scale SY table support to scale the visibilities and the weights by the contents of the SY table. Moved nowhere. 13323. October 24, 2010 CookBook Eric Put back the current draft of Appendix E (EVLA), $DOCTXT/LaCookE.tex. It includes a section on SysPower data. Moved nowhere. 13324. October 24, 2010 ATLOD Eric Michael Bietenholz has found yet another thing that caused aborts. Why did these not cause issues when we tested them? The call to JPRECS requires a variable for the EQUINOX, not a numeric constant. Moved nowhere. 13325. October 25, 2010 SNEDT Eric The application of the FG table to the SY, TY, etc table read in did not trigger an update of the table on disk all by itself - which it should have. Added 1 line to QEDIUTIL and moved a number of places where thee change counter was initialized to before the call to EDFGAP. Moved nowhere. 13326. October 26, 2010 DOCRT Eric DOCRT > 0 is supposed to trigger CRT printing and an interactive task. Discovered that some programs were requiring it to be > 0.5 or they would resume AIPS. Some of these would then go into interactive printing anyway, leaving the two programs fighting over the terminal window. Fixed: INPUT OOP routine INSTRT traps DOCRT - made it get the real float value and then set either integer or float as the task desires with a correct interactivity setting. This fixes tasks BSPRT, FINDR, FLAGR, RFI but they were not bad ones on odd values of DOCRT. LISTR One of the bad ones - fixed RQUICK test. RLDIF One of the bad ones - fixed RQUICK test and setting of IDOCRT. SHOUV One of the bad ones - fixed RQUICK test. BLSUM RQUICK not controlled by DOCRT at all - bad no matter what. Moved nowhere. 13327. October 27-28, 2010 SY tables Eric Changed GETCDS New subroutine to read all Tcals - flags <= 0 and reports them. SNPLT OPTYPE PSPD -> PSYS and plots Tsys = Psum/Pdif/2*Tcal ; also help. QEDIUTIL Changed the SY routines to call GETCDS and to convert the Psum/Pdif/2 to a real Tsys. Fixed labeling and units. SNEDT Changed help and HLPEDISS EDITA Changed help and HLPEDISY TYAPL Changed to use GETCDS and to allow more than 2 IFs (which was in to allow debugging). Moved nowhere. 13328. November 11, 2010 TYAPL Amy For the situation where there was data on a new source with a time stamp before the first entries in the SY table the subroutine GETSY in TYAPL would read the entire SY table looking for bounding times, not finding any. Then since the table had been read it would not be able to find bounding times for any time thereafter and all the weights would be set negative, effectively flagging the data. I fixed GETSY in an inelegant way, and this change should probably be propigated to the subroutine that reads TY tables. Moved nowhere. 13329. November 24, 2010 DATCAL, CGASET Eric Have CGASET get the delay as a delay in phase units rather than as cos(delay) and sin(delay). Then DATCAL can know the full delay difference per channel. This enters into a per channel complex multiplier for walking through the channels but can be used with full dynamic range for the offset from the reference channel to the first channel. Thus very large delay corrections will be handled. Changed it to use CURCAL(3) for delay and (4) for rate. The fifth variable now has no use... Moved nowhere. 13330. November 24, 2010 BDF2AIPS Amy/Eric Added new parameter SWOrder with a question to choose to reorder the IFs to strict frequency order or by default leave them in the order of the data set. Moved to /home/AIPS also. 13331. November 24. 2010 DADEVS.PL Eric The @ sign was missing 3 places where it should have been. Newer Perl versions have started to care. Moved nowhere. 13332. November 25, 2010 CDSEL Eric New subroutine to select IFs etc when copying CD tables. Changed COPTAB, UVCOP, SPLAT to use it. Moved nowhere. 13333. November 25, 2010 VBGLU Eric Added SY, CD, PD, and CP tables to the IF gluing tables subroutine. Moved nowhere. 13334. November 26, 2010 Info files Eric Updated tab completion file HLPIT.LIS and apropos data file LSAPROPO.HLP using hlpitgen.pl and HLPA tools. Using SHOPH updated ABOT files: ADVERB ANALYSIS AP BATCH CALIBRAT CATALOG EDITING GENERAL HARDCOPY IMAGING INFORMAT INTERACT MODELING OOP PLOT POLARIZA RUN SPECTRAL TABLE TASK TV-APPL TV UTILITY UV VERB VLA VLBI Moved nowhere. 13335. November 26, 2010 SHOPH Eric Limit category names to 8 characters automatically. Got into trouble with spellings differing after 8 for same category. Moved nowhere. 13336. November 29, 2010 CookBook update Eric Changed chapters: 0 Dates, T of C 1 Mention Appendix E 2 Changed DADEVS structure 4 FITTP, TYAPL etc (SY table), TVFLG labels, POSSM vector avg, CALIB averaging/normalization, SETJY 2010 values and spectral index, EVASN, BPASS normalization & spectral index, SOUSP, PCAL spectral ref Appen E 5 CALIB normalization 6 SNPLT SY table, POSSM/BPLOT BD, PD, CP tables, TVHELIX, RSPEC 7 CONVL added to resolution, RM added 8 NOIFS, IMAGR/MCUBE/CONVL units 9 RLDLY replaces VLBACPOL, FRING CHINC, halves, ACCOR usage statement, FITLD ANTNAME, UVAVG options 12 Remote batch, ADDDISK E UVAVG OPCODE, POSSM 0 -> vector now I, J Update indexes Recipe rearranged as needed. The html and pdf files were also updated. Moved nowhere. 13337. November 29, 2010 ZTACTQ Eric Changed $APLUNIX/ZTACT2.C to include the errno when a stat on the /tmp/task.nnnn files gets an error. AIPS.EXE affected. Moved nowhere. 13338. November 30, 2010 TYAPL Eric Reviewed Amy's changes and extended them to the TY subroutine with minor typing changes. Moved nowhere. *********************************************************************** *********************************************************************** Changes while 31DEC10 is TST *********************************************************************** *********************************************************************** 13340. December 1, 2010 RMSD Eric The MACPPC compiler cannot handle a parameter that involves > 1 math operation. Thus PARAMETER (A = B*B/32) does not work whilst PARAMETER (A = (B)*(B/32)) does. Moved to 31DEC10 this date. 13345. December 3, 2010 SNEDT Eric A badly typed IF statement caused a computation of difference values with EDIFSY (for SY tables) when SHOW ALSO RATE was selected in editing SN tables (and probably in editing UV data). Fixed it to do only SY there in QEDIUTIL. Moved to 31DEC10 this date. 13347. December 6, 2010 LISTR, ANFIND Eric LISTR set 2 READ statuses when OPTYPE not recognized. This messes up ANREFM which clears one only. Changed ANFIND to recognize SAO SMA array. Moved to 31DEC10 this date. 13349. December 7, 2010 VLAPROCS, VLATECR Amy Added a message to VLATECR about the possibilty of the download failing if the VLA project is too recent. Moved to 31DEC10 Dec 8. 13350. December 8, 2010 FITLD Eric FITLD had trouble with a FITS IDI table that expressed a logical variable as a character-string valued variable. Changed GETLG to ignore an open quote mark but then require the value to follow immediately. Moved to 31DEC10 this date. 13352. December 9, 2010 TVFLG Eric Added option LOAD NEXT CH/IF/ST which increments the spectral channel and reloads the TV. If spectral channel exceeds the max, it is reset to BCHAN and the IF is incremented. If the IF exceeds the maximum, it is set to BIF and the polarization is switched possibly with a Stokes flag change. Corrected the help file and HLPTVFLG.HLP, the run-time help file. Moved to 31DEC10 this date. 13353. December 10, 2010 TYSMO Eric All other SY tasks changed to use Tsys rather than Tsys/Tcal so I changed this task too. In so doing, I found a number of issues with the handling of SY clipping - mostly because TY has only one real column and SY has tow columns and a combination and all are related to each other. These include (1) clipping Pdif and Psum before using them to compute Tsys, (2) blanking Pdif, Psum, Tsys if Pdif <= 0 or Psum < Pdif or either Psum of Pdif blanked (e.g. by clipping) (3) blanking Pdif and Psum if Tsys outside range allowed (4) rearrange logic in testing result after MWF smoothing Moved to 31DEC10 this date. 13357. December 16, 2010 SOUSP Eric Added CALCODE to the adverbs so that identical SOURCE/QUAL entries in the SU table may be differentiated. Moved to 31DEC10 December 28.. 13360. December 22, 2010 TRANS Eric The new TRARow algorithm to figure out number of planes did not take into account the fact the images have an integer number of rows per disk block or an inetegr number of disk blocks per row. Thus if NX = 401, the actual space used is as if NX = 512. So it thought it could do more than it could and aborted. Moved to 31DEC10 this date. 13361. December 24, 2010 CALIB Eric The subroutine that applied gain normalization to the SN table on an IF/polarization dependent basis messed up by trying to use the same buffer for read and write. Switched to 2 LUNs and buffers with TABSN and it still did not work. Changed to straight TABIBI and TABIO and that does work. Moved GMODIT routine to 31DEC10 this date. 13366. December 28, 2010 SOUSP Eric SOUSP did not check for fluxes <= 0 in some IFs. This led to NaNs and so I fixed it. Moved to 31DEC10 this date. 13367. December 29, 2010 DBCON Eric Added messages on appending tables and added SY table to list of those appended. Moved to 31DEC10 this date. 13368. December 29, 2010 SNEDT, EDITA Eric Plot scaling for SY tables was affected by an error in the coding. Moved the minimum plot scaling values to a larger array and set them in the individual data type main routines in QEDIUTIL. This means the e-13 scaling for rates is not affecting Pdif/, etc. Moved to 31DEC10 this date. 13371. December 30, 2010 PCAL Eric The task kept track of source numbers as SID(source#)=cal# but then used SID in a form that should have been CID(cal#)=source#. Added the alternative list so that internal use and use with the CP table would work. CP table ended up with wrong values for source number. Moved to 31DEC10 this date. 13372. December 30, 2010 RLDIF Eric The corection to the PD table used the same sign when correcting the R and L sides. The hermitian should be used for L polarization. Moved to 31DEC10 this date. 13373. December 31, 2010 DBCON Eric Added CALCODE to the tests of how to renumber sources - it was losing some of the distinctions that the EVLA has put in. Moved CALCODE correction to 31DEC10 this date, not NX one. 13375. December 31, 2010 RLDLY Eric An array was dimensioned MAXANT and was used as if MXBASE which walked on things. Fixed it - why did it seem to work? Moved to 31DEC10 this date. 13376. January 3, 2011 LISTR Eric The GAIN listing applied any gain renormalization prior to testing for magic blank. That made magic blanks not be magic and so messed up the listing. Gain normalization application in LISTR was dated Dec 8, 2010 and so is now added to 31DEC10. Moved to 31DEC10 this date. 13380. January 6, 2011 RLDLY Eric Found an error handling the DOIFS=2 option and probably also the DOIFS=1 method. There was a loop over antennas that should not have been run when spreading the results into all of the IFs except for 1 antenna each time. Moved to 31DEC10 this date. 13382. January 7, 2010 FRING Eric 1. FRQTAB needs to do all frequency computation in double precision. With a small increment and large center frequency the resulting list was not accurate enough for FRING. 2. FRING failed to test after QINIT that it got all the memory it wanted. Added tests and messages about SETMAXAP. Moved to patch January 21. 13384. January 18, 2011 DBCON Eric Changed DBCON to set NREC which should defend it against bad BUFFER(42) values. Moved DBCON NREC value only to patch this date. 13385. January 18, 2011 UVCOP Eric Used only the first 12 of the 16 characters of the source names. EVLA tends to be profligate so this no longer works. Moved to patch this date. 13386. January 18, 2011 CALIB Eric The header for the output file was not previously prepared by UVGET although the data for the file were. This made it possible for a FREQID random parameter to appear in the output file with completely uninitialized values. Changed the file creation to call UVGET and then modify the header as needed to put back compression if the output was to be compressed. Moved nowhere - moved to patch 24 January. 13388. January 18, 2011 RLDIF Eric 2. Improve help file description of the task - it was obsolete. 5. Found that the one-source, continuum method in which the SU, AN, and CL tables are corrected did not work. Required repairs to the addressing. Moved these parts to patch 24 January. 13395. January 20, 2011 OOP Eric Subroutine OUVATT in UVDATA.FOR searches for the catalog number of the file after it is created. For some erroneous reason it searched for ISEQ=0 meaning to get the highest SEQ not the one that we were creating. It then attached the catalog number of the highest SEQ to the object. The next time the file had to be found, the wrong one was found. This caused SPECR to copy the tables to the highest SEQ file rather than the one it had just filled with data unless that one happened to be the highest SEQ. With different code sequences it could cause the new file itself to never be filled (except with 0's) and the old higher SEQ file to be overwritten! Moved to patch January 21, many OOP tasks must be relinked. 13398. January 21, 2011 FRING Eric The exhaustive search method suffered a serious error in which one of the parameters giving the number of channels per IF was divided by the number of IFs a second time. This led to correct answers for the first IF but wrong ones for subsequent IFs. Moved to patch this date. 13401. January 24, 2011 SWAPR Eric Ignored the DOACOR adverb - fixed to put that in common for UVGET. Moved to patch this date. 13450. March 23, 2011 SNPLT Eric SNPLT failed to handle PC table phases correctly because at one critical point, the data type code is changed to special values just for PC tables. When other operations on SY tables were added, these special values were not updated at every place that they were used. Phase is in fact handle carefully to try to determine if -180 to 180 or 0 to 360 plotting would be better and the better one is chosen. Moved real correction to patch this date. 13481. April 15, 2011 NX table writing Eric RNXUPD wrote the new subarray number in the NX table when dumping records for other subarrays, i.e. when switching from subarray 1 to subarray 2. This messes up CALIB and causes it to omit the last scan of subarray 1 in self-cal. Fixed local copy of RNXUPD in SPLAT also. Moved to patch this date. 13482. April 15, 2011 IMAGR Eric Had an error addressing multiple channels when doing time averaging in a baseline-length dependent way. Led to smears of signal going out radially. Moved to patch this date. 13526. June 7, 2011 IMEAN Eric DROUND was declared double precision when it returns integer. This seemingly worked under GNU but not under Intel! Moved to patch June 8. 13462. March 31, 2011 NOIFS Eric Found an error in setting the UVW scaling parameter. Moved nowhere - moved to patch June 14. 13539. June 23, 2011 CVEL Eric Programmers got confused in the init subroutine about which variable actually went back to the calling program as the error code. The wrong one was used for many important things like BP file setup, causing the task to plow on in numerous hopeless circumstances. Moved to patch this date. 13541. June 24, 2011 FITS Table read error Eric The subroutine ATCONV adjusts the list of table keywords for a couple of effects including the presence of ISORTORD. It did not adjust the list of character valued keywords so they no longer corresponded to the list of types, names, pointers, etc. Fixed ATCONV and changed the call sequence in FITLD, UVLOD, IMLOD, CALRD, and TBIN. Wht was this not noticed? Moved to patch this date. 13497. May 4, 2011 FITLD Eric Changed the local parameter include to contain more of the parameters and put that in SUCHCK rather than declaring MAXSOU as only 1000 while it was 2500 elsewhere. Changed it to 5000 at the same time. Moved nowhere, to patch June 24. 13460. July 1, 2011 UVCOP Eric In order to flag SY, TY, SN tables changed UVCOP Changed NXTFLG - the SOUWAN list is prepared differently by UVCOP and is always sources that are wanted. The usual flag about that said otherwise. This meant that a UVCOP with SOURCES did not apply any flags to UV data or to table data. Moved UVCOP fix to its NXTFLG to patch this date. 13490. July 25, 2011 IMAGR Eric Changed QCLEAN: 1. CLSCAL will now write a message whenever rescaling by something other than 1.0. Previously, if the scale was exactly the previous one it siad nothing. 3. CLGFIT now uses the field for which a beam was computed most recently to provide the beam parameters. Previously it used each field even if ONEBEAM caused that field not to have a beam. As a result it read the header beam value and used it for all planes of the image in scaling, restoring, etc. It did write this bad value in the CG table but who looks? The new value does get into all facets from this point. 4. CLSCRS changed to display all rescaling not = 1.0 Moved to patch this date. 13495. July 29, 2011 install.pl Eric Added code to recognize 64-bit Macs as MACINT. Moved to 31DEC10 this date (not patch). 13496. July 29, 2011 VPLOT Eric The routine that picked up the CC models had an error if NMAPS was greater than 64. A wrong variable was used in an IF statement so NCOMPS for field 66 was set to NMAPS and so all other fields were ignored. Moved bug fix to patch. 13502. August 4, 2011 CLCOR Eric Two significant places - opcodes 'EOPS' and 'SUND' - assumed that the antenna coordinate system is left handed. These did NOT get corrected in the fall of 2009 as they should have. Moved to 31DEC10 this date.