;----------------------------------------------------------------------- ; Copyright (C) 2013-2015 ; 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 31DEC14 is TST *********************************************************************** *********************************************************************** 14257. December 1, 2013 Version update Eric Changed $APLSUB/GETRLS.FOR. $INSUNIX/INSTEP1, $SYSPERL/install.pl, $SYSPERL/update.pl, $SYSPERL/ZXHLP2.PL, $ SYSPERL/LaCook.PL, $UPDUNIX/UPDCONTROL, $SYSUNIX/AIPSPATH.CSH, $SYSUNIX/AIPSPATH.SH to make 31DEC14 instead of 31DEC13 - also fix dates in install.pl. Moved nowhere. 14258. December 2, 2013 DOFARS Eric The procedure had not been changed when FARS took over control of its output class names. The final TRANS tasks then failed. Moved to 31DEC13 this date. 14259. December 4, 2013 VLAMP Eric Changed it to compute central frequencies at channel N/2.0 + 1.0 and added comment to say what it is doing. Moved nowhere. 14260. December 5, 2013 APCAL Eric Changed GC reading to understand that a GC subarray of 0 means all. Changed table reading in general to understand that a negative error return means flagged row, not severe error. Moved to 31DEC13 this date. 14261. December 5, 2013 USUBA Eric Changed the subroutine that automatically changed the subarray column value to use its subarray magic only on tables with subarray, source, time, and antenna columns and to output subarray = 0 for other sorts of tables. Subarray 0 means any which many tasks now understand. Moved to 31DEC13 this date. 14262. December 5, 2013 APCAL Eric Changed it to loop over subarray if SUBARRAY <= 0. It meant dealing with the insistence on deleting the old SN if it existed and then starting at the beginning (was overkill). Changed VLBACALA help file also. Moved nowhere. 14263. December 6, 2013 SUBIM Eric The YINC=1, ZINC=1 code was broken... by the recent upgrade. Moved to 31DEC13 this date. 14264. December 6, 2013 AIPS Eric When an error in POPS occurs or a task returns a failure code, OERROR is called and control returns to the top most level of interactive input. OERROR did not close nested RUN files properly or reset NUMRUN = 0. Fixed this. Moved to 31DEC13 this date. 14265. December 11, 2013 HELP Eric Added help paragraphs about the CookBook and where to get the PostScript, pdf, and html versions. Moved to 31DEC13 this date. 14266. December 11, 2013 CookBook Eric Changed COOKNS and COOKNX to add a new appendix (L for low frequency). Moved nowhere. 14267. December 11, 2013 HIIO Eric HIIO changed the history file header even when the file expansion failed. This could cause the system to think the HI file is larger than it is, leading to persistant problems. Moved nowhere. 14268. December 13, 2013 TGET file Eric Changed the format of the TG file and its name - the adverbs will now be stored with 3 words ahead of them, 2 for the adverb name and 1 for its length and type. Then a TGET will only change adverb values for those adverbs found in the file rather than being messed up by the inclusion of a new adverb. Changed: AU2 Changed code to build the TGET file data in a separate buffer with separate pointers and including the adverb names and adverb size/type. The TG file now gets 8 blocks per task (was 4) and the directory now gets 10 records (was 6). The TELL and task-data files still get 4 per task. AU2A Added code to do TGET from the new format TG file and then to compare the adverbs found there with the current INPUTS file. Also TGINDEX will do first the new format file and then the old, listing from the old only tasks not found in the new. AU3A Updated TIMDEST code even though it is blocked. Corrected the code that deleted the old format task save file to delete the more modern format ones on SAVDEST. MV2C06TG The old MV2C06TS file in $DOCTXT was renamed and edited to bring it up to date, poiting out the differences between the now current format and that which was used for quite afew years previously. Moved nowhere. 14269. December 17, 2013 CVEL Eric VERA apparently follows the lead of the VLBA and fringe rotates to the Earth center. Change ISVLBA to detect array name VERA as well as VLBA and generalize the messages about that. Moved to 31DEC13 this date. 14270. December 18, 2013 RLDLY Eric Changed code around so that REFANT=0 means loop over all possible REFANTs. The top level code does a weighted average of all the solutions and applies that after the loop over antenna. Moved nowhere. 14271. December 19, 2013 BLSUM Eric Values of TXINC and TYINC of zero were used rather than being corrected to the default value 1. This led to aborts, infinite loops, etc. Moved to 31DEC13 this date. 14272. December 20, 2013 CookBook Eric Add new Appendix L for low-frequency EVLA reductions in AIPS. This makes changes to chapters 0, F, I, J also. Changed Recipe.tex and LaCook.mac (for figure macro in use for some time). Moved to ftp site, nowhere else (but goes to all versions). 14273. December 24, 2013 SLFIT Eric The conversion to REAL*8 in lots of places on July 8 was too enthusiastic and sent R*8's into a routine that requires R*4's. The conversion from fit parameters to real units for most slices produced wrong answers. Moved to 31DEC13 this date. 14274. December 24, 2013 Maximum image dimension Eric Changed PMAD.INC parameter MAXIMG to match MAXCHA in PUVD.INC at 131072. I hope no one tries FFTs of this size but we must be able to build images with the maximum number of spactral channels. Moved nowhere. 14275. December 25, 2013 AIPSLetter Eric Put the 31DEC13 AIPSLetter as LET13B .TEX, .PS, and .PDF into $AIPSPUBL and on the ftp site. Moved nowhere. 14276. December 26, 2013 smoothing Eric The 2-point smoothing routine TPTBSM tried to avoid a divide by zero by taking the max of 0.001 (86.4 seconds) and the time difference of the 2 points. Since they might be around 1 second this is very bad. Fixed the code to avoid a zero divide explicitly or use whatever the time difference might be. Affects CLCAL, SNSMO, TYSMO, CLSMO, and SNEDT. Moved to 31DEC13 this date. 14277. December 27, 2013 BLSUM Eric Changed: 1. Modernized the code including making the mask be in dynamic memory rather than on a scratch disk. 2. Corrected the labeling - when the blotch was, e.g. a moment image, then the labeling of the spectrum implied it was in Jy/beam when in fact it was in Jy. 3. Corrected the blotch setting to make sure that each vertex is in the image. 4. Added the option to plot the blotch spectra in real plot files, with 1, 2, 3, or 4 spectra per page. 5. Corrected the printed and plotted channel numbers to account for BLC. Moved nowhere. 14278. January 13, 2014 PCAL Eric Found two places that referred to CATIN using UV header pointers that had been recomputed for the header after calibration. Old data sets might have a no longer standard axis order and so the two might differ. This caused it to lose an IF when saving the results in a pure continuum case. Moved tp 31DEC13 8-Aug. 14279. January 13, 2014 DBCON Eric One delta angle for source position comparison was much larger than claimed in the comment - fixed it to 10 milli-asec. Moved nowhere. 14280. January 13, 2014 IMAGR Eric Added the concept of UNClean boxes (areas where components are never found). This is a way to block out an area in one facet and have that area Cleaned only in some other facet (to be used later in modeling for example). Changed: TVUTIL $APLOOP: added concept of UNClean windows in OTVUBX, made sure that the pointer was zero everywhere Clean windows pointer was zeroed, made sure that all TV loading called OTVUBX if the pointer is not zero. QCLEAN $QOOP: added Common variables ISUNBX, UNBOXS, and UNWIN to indicate if UNboxes are used and if so how many and their values. Pick these up in the INI routine and add them to routines BOXFIX and WRIBOX. Changed the TV curvalue to use graphics plane 2 so that UNClean boxes use plane 4 (cyan). Added UNBOX to the menu when ISUNBX is true. Added UN-boxes to the auto-boxing, histogram, pixel finding, SDI Clean, and maximum finding routines. Added local version of WRIBOX to write UNClean boxes and to report errors more carefully. Added routines UNBROW and UNBCOL to aid in applying UNClean boxes. IMAGR $QPGOOP: drop GRCHAN as an adverb - use channel 5 always. Add and use routine UWINDF to read UNClean boxes from BOXFILE. Then put the UNClean boxes if any in the Clean object. IMAGR $HLPFIL: drop GRCHAN and add info about U cards in BOXFILE. BOXFILE $HLPFIL: add same remarks more or less. HLPCLEAN $HLPFIL: add UNBOX and fix the TVBOX wording. Moved nowhere. 14281. January 14, 2014 BPASS Eric When SOLINT=-1 all scans and sources are averaged together. Under this circumstance, if there is more than one source, there should be no spectral index correction. However, the code made no correction even if there was only one source. Fixed it to set the source number appropriately when there is one and to warn when there should be a spectral index correction, but it is zero. Moved to 31DEC13 this date. 14282. January 14, 2014 LISTR Eric The gain scaling routine picked up the LUN of the open CL, SN, etc. file and then used it to try to open the source table! Changed it to get an LUN for the source table dynamically with LUNTMP. Moved to 31DEC13 this date. 14283. January 15, 2014 SNPLT Eric Added OPTYPEs 'POWR' and 'PODB' to plot (1/GainAmp)^2 and -20 * log10 (GainAmp). Note that GainAmp in the CL tables is an inverse voltage gain so the squared here puts things in observed power units. These options are probably mostly to make plots to show the engineering staff. Moved nowhere. 14284. January 16, 2014 IMAGR Eric Fixed a bug in QCLEAN that made it find no residuals to clean when there was a circular box and no UNClean boxes. Moved nowhere; updated immediately. 14285. January 16, 2014 DBAPP Eric NEW TASK: Appends one or more similar data sets to a pre-existing one. It can re-number sources, but will not do the fancy things with antennas, subarrays, frequency IDs and the like that DBCON can do. It does avoid the 2^N file proliferation problems encountered by DBCON however. Moved nowhere. 14286. January 22, 2014 DBAPP Eric Corrected a test that used the number of axes when it should have used the number of random parameters. Moved nowhere. 14287. January 23, 2014 IMAGR Eric Fixed QCLEAN - the TV menu was self-destructive when there was no UNClean boxes. Moved nowhere, updated immediately. 14288. January 23, 2014 REWAY Eric When data in a multi-source file scan had fewer valid points than allowed by BPARM(1), no record was written to the WT file. The read back to apply these weights then failed because the WT file had too few records (assuming that the bad scans were at the end). Changed to write the WT file when any data were present even if it is writing flags for the too little data scans. Also changed all routines to write out a vis record only if it contains some valid data. Moved nowhere. 14289. January 24, 2014 FITLD Eric Put in special code totrap a new keyword for VLITE named PRIBAND. Added code to save the CORRELAT and PRIBAND keywords in the output header. They go first to the scratch file and then must be copied to the output, with special scratch file keywords omitted. Moved nowhere. 14290. January 27, 2014 MODIM, RMFIT Eric Added RM thickness to the models and model fitting. Three different mathematical forms may be used - sin(x)/x, Gaussian, and exponential. It really seems to solve for the depolarization. Moved nowhere. 14291. January 28, 2014 DISKU Eric Changed DOALL to be the minimum size for individual file listing (in Mbytes) and changed the appearance of the file listing to give file catalog number, type, name, class, seq, size, and sum of extension size. Should help identify large files for managing disk space. Moved nowhere. 14292. January 29, 2014 DISKU Eric Added a choice to use the printer or DOCRT > 0 instead. Will be useful if making long lists of large files. Moved nowhere. 14293. February 1, 2014 FARS, TARS Eric Changed: TARS To compute the Gaussian convolution function only when useful and keep track of that range. Rearranged the convolution code to do, in most cases, far less work. FARS To compute the Gaussian convolution function only when useful and keep track of that range. Rearranged the convolution code to do, in most cases, far less work. Rearranged the component subtraction (inner Clean) code to simplify (like TARS) which made it work - the Clean was not doing the right thing somehow (as if smoothing with a very narrow Gaussian instead of the correct one) previously. Added ability to print some Cleans. FARS $HLPFIL: added PRTLEV. Moved nowhere. 14294. February 1, 2014 RMFIT, XGAUS, ZEMAN Eric Changed: POPSDAT Create new adverb RMSLIMIT - also NEWPARMS.001 and a help file for the adverb. RMFIT Fixed the phase to be between -90 and 90 in the PARMS and thereby in the RESULT and in the display for the user. Changed the CURVAL display to use the actual image values rather than the scaled values from the TV memory. Add RMSLIMIT adverb to help and Fortran to specify an RMS which causes the task to stop automatic fitting and return to interactive. XGAUS Changed the CURVAL display to use the actual image values rather than the scaled values from the TV memory. Add RMSLIMIT adverb to help and Fortran to specify an RMS which causes the task to stop automatic fitting and return to interactive. ZEMAN Changed the CURVAL display to use the actual image values rather than the scaled values from the TV memory. Add RMSLIMIT adverb to help and Fortran to specify an RMS which causes the task to stop automatic fitting and return to interactive. Add display of RMS so that RMSLIMIT can be set meaningfully. Moved nowhere. 14295. February 3, 2014 ZEMAN, XGAUS, RMFIT Eric ZEMAN Fortran: set the number of "Gaussians" to zero when doing the spectrum-average field which made it not actually fit that field. ZEMAN Help: re-wrote the explain file which was never edited for ZEMAN. It was basically for XGAUS before. Added a note about re-starts but they are different in ZEMAN than in the other 2. XGAUS Help: added notes about restarting with various values of NGAUS to simplify fitting sub-regions. RMFIT Help: added notes about restarting with various values of NGAUS to simplify fitting sub-regions. Moved nowhere. 14296. February 3, 2014 PCCOR Eric Fixed usage where it assumed that the frequencies of the pulse tones in R and L are identical. Changed it to use the polarization-dependent values even though they are likely to be the same. Moved nowhere. 14297. February 5, 2014 XGAUS, RMFIT, ZEMAN Eric Changed all 3 tasks so that DOTV=2 says to ask the questions about re-guessing the initial values, marking bad, etc are done with a TV menu instead of the queries in the AIPS terminal window (which are still there on DOTV 1). Changed all 3 Fortran files, 3 help files, and HLPXGAUS, HLPRMFIT, and HLPZEMAN also. This required a change to the call sequence to TVMENU, adding a horizontal offset from the TV edges. That change forced a corresponding change to TVDMEN in TVDEVICE plus call changes in QEDIUTIL, QCLEAN, and tasks TVHUI, TVRGB, WIPER, FILIT, PLAYR, SCIMG, and SCMAP. Moved nowhere. 14298. February 6, 2014 RMFIT, ZEMAN Eric Changed RMFIT to plot the initial guess for a component that will not be fit to the side of the plot rather than at zero. This will let users see that it is nulled rather than simply zero. Added option SET MIN P1 to revisit all weak pixels if one desires. Changed help and real-time help for this too. Changed ZEMAN to remove bugs that kept the TV turned off on DOTV=2 and caused an abort when it did call for the second new menu. Moved nowhere. 14299. February 7, 2014 UVFND Eric Added the DOSCALE option - to ormalize the fluxes by the fluxes in the SU table - including spectral index if desired. In testing it was found that channel averaging had a bad bug - the imaginary part of the output was the same as the real part giving a phase of 45 degrees and an amplitude sqrt(2) too high (or so). This was not done when NCHAV <= 1. Moved the fix part to 31DEC13 this date. 14300. February 8, 2014 DOBAND 2 and 4 Eric The code of BPGET assumed that the BP correction array was filled in with the earlier of the 2 times. This was not correct at the start and, if the first time(s) were closer to the BP record of the lower time, then the BP correction array was not completed properly. The values must have somehow been approximately correct rather than 0, from earlier work with the array. Changed it to copy the early solution to the array or the later solution to the array in all cases. Moved nowhere - should this be a patch? 14301. February 10, 2014 DOOSRO.001 Lorant Previously I removed a lot of clrtemps since they started to be differently interpreted, I think due to a change in an upper routine in 2013. However, one critical procedure (numtab) actually needed a clrtemp within a loop, which is now there. Added plenty of lines of debugging code to find it. Moved to 31DEC13 this date. 14302. February 10, 2014 FITLD Eric Changed it so that an error in reading trailing records in the input file does not stop the task from completing things - although such an error suggests that there will be issues. Moved nowhere. 14303. February 11, 2014 KNTR, PCNTR Eric Changed: KNTR A test when doing 3-color pictures required BLC(3) < TRC(3) when the setup to that point had them equal. Changed test for axis 3. PCNTR Fixed assignment of LUNs - the 3 color used LUNs = ILUN(4) + 1 and 2. But there is now an ILUN(5) so changed the (4) to (5). TVRGB Corrected call sequence to TVMENU - the ue of constants confused me TVHUI Corrected call sequence to TVMENU Moved KNTR, PCNTR to 31DEC13, others not needed. 14304. February 11, 2014 PCCOR Eric PCCOR did not expect magic blank values for the cable-cal. But the tables have them (double precision magic blank is 10^98) so trouble ensued. Changed to use the last good value of cable-cal when blanked. Moved to 31DEC13 this date. 14305. February 22, 2014 SNPLT Eric Rates are scaled by frequency to mHz for plotting. The task used the frequency of BIF for all IFs. Changed it to use the correct frequency for each IF. Moved nowhere. 14306. February 24, 2014 POSSM Eric Improved the labeling of plots, particularly autocorrelation spectra. Moved nowhere. 14307. February 25, 2014 POSSM Eric Changed labeling some more - set up a 3-state integer to tell whether data are all parallel polarizations, cross polarizations, or mixed and then have 3 separate labels. Added code to convert STOKES = 'CROS' early enough for it to be used correctly (as 'XYYX', 'RLLR', 'QU'). Moved nowhere. 14308. February 25, 2014 SNPLT Eric Corrected the rate difference to apply the frequency correction with the correct IF to the separate rates before the differencing. Moved nowhere. 14309. February 25, 2014 AVSPC Amy In one call of CHNDAT, AVSPC sent a FREQID=0 always. This only caused a problem if the input data set had more than one FREQID. The real FREQID was sent to CHNDAT which fixed the problem. Moved nowhere. 14310. February 25, 2014 CHNDAT Eric The fundamental subroutine CHNDAT meant to read/write FQ tables had a path through it which would leave the LUN open. Corrected this to always make sure it is closed. Moved nowhere. 14311. February 26, 2014 DOOSRO Lorant When VLANT runs it may create a new CL table. When it is skipped or doesn't need to create a new CL table, CL1 gets copied to CL2 to have the same number of CLs hanging around for outclass UVLANT. When reruns happen, it wasn't checking for the possible clash. Included the check now. Moved nowhere. 14312. February 27, 2014 SNSMO Eric Changed: 1. Pass FREQS into CLPRAT and scale each rate by the appropriate FREQS before summing into the work array and then when comparing to the clip level. Previously it used unscaled rates (so in sec/sec) and compared them to a clip level converted to sec/sec using the header frequency. Some modes added together multiple IFs not converted back to Hz (and so having different values solely due to frequency). 2. In CLPAPH, changed it to find the FREQ scaled rate using the frequency of the IF from which it took the rate. Previously it used the frequency of IFBEG. 3. In REFRAT, summed the rates in Hz using the correct frequencies for each IF rather than in sec/sec and made the reverse correction in putting the rates back. 4. Changed the delay smoothing routine into one for multi-band delay and one for each IF. The latter gets pointers to the real and imaginary columns so that the phase may be corrected. 5. Added NPIECE adverb to help and code. It is used to undo the delay phase correction made by FRING and redo it with the smoothed delays. Added frequency difference array to account for the FRING scaling and the difference between the reference pixel and the center of each IF. 6. SMORAT is given the frequency array so that it can sum the rates in Hz rather than sec/sec and then put them back correctly fixing the units back to sec/sec. 7. In SMOAPH, pick up rates in Hz at the frequency of the IF from which the rate info is taken. Correct the multi-band phase term which had the wrong sign for the second polarization. This made R-L phases appear where none were present before. 8. Correct spelling of CUTOFF in history file. Moved nowhere. 14313. February 27, 2014 SNPLT Eric The IF frequency for the second polarization, when differencing polarizations, was not set properly for rates. This led the rate plot with STOKES='DIFF' to be the same as with 'RR'. Moved nowhere. 14314. March 3, 2014 SPECR Eric Added option to interpolate by normal interpolation functions including 2 point and Median window, although the help still recommends the FFT approach. Moved nowhere. 14315. March 3, 2014 FTFLG Eric Filled empty space with zeros rather than magic blanks in the fast gridding routine. Like the fix June 21, 2013 in SPFLG. Moved nowhere. 14316. March 10, 2014 FITLD Eric The subroutine GETMC could fail to do the correct hash interpretation and so fail to find MC records that it wanted and which were available. MC records separated by < 2 minutes triggered this issue. Rearranged so that it searches for the closest MC record. Also changed GETDEL which would have failed if the records were separated by <= 1 interval. Moved to 31DEC13 patch this date. 14317. March 11, 2014 RLCAL Eric Corrected computing of time intervals based on SOLINT, which was basically all wrong. Stopped allowing the Q and U models to be scaled by a flux in the SU table. Made the task pay attention to NMAPS, which must be > 0 to use IN2NAME et al as a model. Moved nowhere. 14318. March 11, 2014 PRTAB Eric When a floating format was requested but an exponential was required, the task used 1 too few characters, dropping the essential digit of the exponent. Moved nowhere. 14319. March 14, 2014 BLSUM Eric The 31DEC14 upgrade to bLSUM got rid of an extra buffer, but the TVLOAD code was still told that the buffer was available. Fixed the task to tell it the correct current size of the IO buffer. It only failed with really large images. Also cleaned up the coding in TVLOAD to a more modern style. Moved nowhere. 14320. March 14, 2014 CURVALUE Eric Gave the CURVALUE verb 2 adverbs. GRCHAN to control which channel is used and NCHAN to control which image plane memory is used (0 -> lowest channel on at current pixel). Changed CURVALUE,HLP, TVALUE (call sequence), TVDEVICE (call TVALUE with 0 for limiting channel), AU6B (pick up adverbs and use in call to TVALUE). The OOP use of CURVALUE then does not need to change, nor does the XGAUS et al use (they have special versions of TVALUE anyway). Moved nowhere. 14321. March 17, 2014 DSKEW Eric The routine in INTERPLATE that implements this algorithm took the SIN and COS of an angle in degrees rather than radians. This made for a rather bad result whan that angle was not zero. Added comment to help file warning about higher level interpolations "overshooting" the input values. Moved interplate to 31DEC13 this date. 14322. March 17, 2014 SHOW and TELL Eric When AU2 was changed to make a TG file with adverb names as well as values, I did not realize that TELL used this code too as a "TG" file rather than a "TD" file. Fixed AU2 for TELL to write the TC file like a TD file (no adverb names). Moved nowhere (this is a 31DEC14 error only). 14323. March 17, 2014 BPWAY Eric NEW TASK: uses RFLAG-like approach to find rms in each channel as a function of time one scan at a time. In each scan, it smooths the channel rmses and then applies the smoothed values as normalized weights. In doing this task noticed and fixed: 1. ZOPEN changed messages to handle LUNs > 99. 2. LUNTMP changed to check both MAP and non-MAP LUNs for use before assigning one. 3. RMEXT changed call to LUNTMP to ask for a normal file LUN rather than a MAP (double-buffer I/O) one. Moved nowhere. 14324. March 18. 2014 CENTR Eric NEW TASK: Changes the frequency reference pixel to Nchan/2+1. This requires re-scaling the u,v,w values so the data set must be copied. Otherwise it would be only a simple header change. This should help with FRING since the phase rocking is really about the center channel not channel 1 which now seems to be the usual reference channel out of the correlators. Changed DBCON so that DOPOS(2,1) > 0 => change the frequency reference pixel to 1 (as was currently done), otherwise change it to Nchan/2+1. Moved nowhere. 14325. March 18, 2014 ANTAB, VLOG Amy Added warning to ANTAB, VLOG and their HLP files about the use of *cal.vlba* files for calibration of RDBE/MARK5C data. *cal.vlba* files are from the VLBA legacy system and do not apply to RDBE/MARK5C data. All VLBA observations starting Feb 1, 2014 are RDBE/MARK5C observations. The warning in ANTAB will issue if the data is from after Jan 31 2014, from the VLBA and not for the VLA. VLOG is designed specically to work on *cal.vlba* files, so the warning simply issues if the data if from after Jan 31 201. Moved nowhere. 14326. March 19, 2014 CLCAL Eric Subroutine SNAPP used a bad subscript to get the subarray and so did not fill in the table of used reference antennas correctly. This then caused CLCAL to fail to re-reference the SN tables as they were supposed to be. Moved to 31DEC13 this date. 14327. March 20, 2014 DOOSRO Lorant Replaced references of getthead in combination with keyword containing 'ext' (in particular numtab) with the new getvers. Moved nowhere. 14328. March 24, 2014 AIPS Memo 117 Eric Revised the wording about coordinates in CC files. They are simple shifts from the reference pixel and must be corrected for image geometry. Changed all 2012 red things to blue, made the 2104 changes red. Moved nowhere. 14329. March 24, 2014 VLAMP Eric Corrected a number of minor issues - the SY table reading begain with record zero which is incorrect. Changed the tests for the amount of extrapolation that is allowed without a warning. Even then the SY table may be flagged while the CL is not, producing extrapolation warnings. Changed the help file to explain this in greater detail. Moved nowhere. 14330. March 25, 2014 SETJY Eric Velocity computations on OPTYPE ' ' or 'SPEC' in error: a. The reference frequency should have been the specified channel (APARM(1)) in reference IF NIF not the reference pixel in NIF. b. Radio formula had incorrect sign. c. Optical formula also had incorrect sign. Moved to 31DEC13 this date. 14331. March 25, 2014 SNP2D Eric Clarified in the code and help that BCHAN and BIF apply to the file identified by INNAME, INCLASS, etc. Added a phase correction to undo the delay correction on an IF by IF basis as an option. Moved nowhere. 14332. March 27, 2014 SNP2D Eric Finally I think we have it right. The output SN table needs to have a delay = line channel phase / line channel frequency where "line" refers to the channel at which the maser provides a useful self-cal solution. But AIPS' DOCAL only corrects phase with delay between each channel and the reference channel. Thus the output SN table here must also have a phase = ref channel freq * delay found above (all with 2*pi and correct signs etc.). Moved to 31DEC13 this date 14333. March 27, 2014 FQCENTER Eric Change POPSDAT and NEWPARMS to create new adverb FQCENTER, change AU2A to default this adverb to +1, and add a new help file. The adverb specifies that the output UV data set should have channel Nchan/2 + 1 as the frequency reference, rather than the usual 1.0. This helps with FRING among other things since the delay phase slope is better represented as +- equal amounts about the average than a value at one end. Added this adverb to the Fortran and Help files for tasks DIFUV DSORC DSTOK REAMP REWGT UVCMP AVER BDAPL BPWAY BPWGT DAYFX DBCON DFQID DQUAL FIXRL FUDGE FXTIM HAFIX NOIFS PHNEG REIFS REWAY RLCOR SWPOL TI2HA TYAPL UVAVG UVCOP UVFIX UVLIN UVLSD UVLSF UVMOD VBGLU VPFLG WTMOD FXPOL MULIF CVEL Also changed VBCAL Added FQCENTER, changed it to allow compressed data, modernized coding some UVMOD Wording in help not quite correct - FPOS is in coordinates not tangent plane. WTMOD Add some clarification in help. Moved nowhere. 14334. March 27, 2014 SPLAT Eric SPLAT, unlike other tasks, "tidies" up the SU table (omits sources which have no data in the output file). It has left the CL table with entries for these sources however. Added a CLTIDY routine to clean up the CL table also. Moved nowhere. 14335. March 31, 2014 TIORD Eric Correct formats of display which also includes the visibility number (missing from FORMAT statement). Clean up wording in help file. Moved Fortran corrections to 31DEC13. 14336. March 31, 2014 BPWAY Eric Changed: 1. Added OPTYPE 'SORC' to do one source at a time rather than one scan at a time. Report when data are out of order and set sort order in output header. 2. Added option to average the Stokes rms values before smoothing. The routine that does this also finds the maximum number of WT file records in any one baseline. 3. This max count is used to make dynamic memory for the smoothing, rather than assuming that any of the current dynamic or fixed buffers is adequate. 4. Make sure that the INDEX arrays do not overflow (die if it does which is very unlikely). 5. Fix header for actual number of vis written and compress file. 6. Fix mistaken use of BUFF2 in the smoothing (it is in use for UV data from previous scan or source). Moved nowhere. 14337. March 31, 2014 INDXIN Eric Subroutine which intializes the NX table in calibration routines could exit leaving the NX table open when there were no lines in the NX table. Then the cal routines would fail with LUN already open in FTAB messages. Moved nowhere. 14338. April 1, 2014 TDEPEND Eric Imaging a time-dependent source in the presence of other time-independent sources is a bit of a problem. Created TDEPEND $HLPFIL: general help about the issue TDEPEND $RUNSYS: RUN file to compile useful procedures TD_SCANS $HLPFIL: procedure to find scan breaks TD_SSCAN $HLPFIL: procedure to find sub-scan breaks TD_STEP3 $HLPFIL: procedure to make initial data set having the target source removed TD_STEP5 $HLPFIL: procedure to make later data sets having the target source removed Moved nowhere. 14339. April 2, 2014 XGAUS, RMFIT Eric Discovered numerus calls to the function evaluator which had a bad argument, causing it to evaluate derivatives rather than residuals. It is amazing that it worked so well - the symptom arose in XGAUS where a single variable (an FTAB pointer) got reste to zero most mysteriously in Intel 64-bit optimized code only. Correcting this unearthed an error in RMFIT's attempt to make guesses - the change from 3 variables to 4 fit each time did not make it to one place in that code. Moved corrections only to 31DEC13 this date. 14340. April 2, 2014 PRTAB Eric A very long string in a table (BANDCODE for 64 IFs = 256 characters) was too long for temporary strings in the task. Limited the data string in such cases. Unfortunately, without building in special knowledge of specific tables, PRTAB cannot know about character string arrays. Moved to 31DEC13 this date. 14341. April 2, 2014 histry Eric This is a tool for AIPS programmers on Dave which lists the putbck history of files. I corrected the script to handle files in $AIPSPUBL and such version-independent ares and also to find the REASONS.UPD files to grp. The latter required some new link files in another area (which must be maintained at rollover time - note added to rollover instruction file). Not part of code movement. 14342. April 3, 2014 VLBAUTIL Amy Corrected messages out of VLBAEOPS, that were no longer correct since INFILE was removed as an input from this proceedure. Added deletion of downloaded file to VLBATECR, which was a problem for the VLBA pipeline. Moved nowhere. 14343. April 3, 2014 documentation Eric Updated HLPIT.LIS $HLPFIL: tab completions for new tasks, adverbs, procs APROPO $HLPFIL/LSAPROPO.HLP - updated the Copyright, the rest is updated by putbck. ABOUT lists $HLPFIL: ADVERB, ANALYSIS, AP, CALIBRAT, IMAGING, INTERACT, MODELING, OOP, PROCEDUR, SPECTRAL, TASK, TV-APPL, UV, VLA Moved nowhere. 14344. April 7, 2014 FQCENTER Eric If the frequency axis is centered, the FQ table may need to be changed if some of the IFs have a different channel separation than that listed in the header. Wrote subroutine CENTFQ to do this correction and changed the following tasks to implement this correction: CENTR DIFUV DSORC DSTOK REAMP REWGT UVCMP AVER BDAPL BPWAY BPWGT DAYFX DFQID DQUAL FIXRL FUDGE FXTIM HAFIX PHNEG REWAY RLCOR SWPOL TI2HA TYAPL UVAVG UVCOP UVFIX UVLIN UVLSD UVLSF UVMOD VBCAL VBGLU VPFLG WTMOD MULIF CVEL Changed FXPOL to drop the FQCENTER adverb - too messy. Tasks DBCON, NOIFS, REIFS, and SPLAT do not need changing here. Moved nowhere. 14345. April 9, 2014 APCAL Eric Changed it to avoid being fooled by duplicate records in the GC table. VLBAMCAL (or VLABFIX) really should be used in cases where this is a problem, but we muight as well fix this here too. Moved nowhere. 14346. April 10, 2014 CookBook Eric Updated as follows, files $DOCTXT/LaCook*.tex, $AIPSPUBL/COOK*.PS 3 TGET file format, DISKU DOALL meaning 5 UNClean boxes, BPWAY, TDEPEND 6 BLSUM makes real plot files 7 BLSUM, RMFIT & MODIM polarization thickness 8 BLSUM, SPECR non-Fourier methods 9 USUBA subarray=0 meaning, RLDLY loop over refants, DBAPP, CENTR and FQCENTER re FRING 11 DISKU (DOALL usage) 13 Numerous ABOUT lists E RLDLY loop over refants, BPWAY, CENTR & FQCENTER, SNP2D Z Overhaul, reduce info re Charlottesville, metion Green Bank 0 updated table of contents, plots, recipes I, J index updated COOK*PS copied to ftp area including gzip copies. Made html version at aips web site. Made and putbck pdf version $AIPSPUBL/COOKBOOK.PDF Moved nowhere. 14347. April 11, 2014 LISTR Eric LISTR failed to display the correct scaling when FACTOR was used with phases, elevations, or parallactic angles. The verbose listing (DPARM(5) > 0) also did not show when numbers overflowed the format. The help file needed clarification on the use of FACTOR (which is always used in GAIN listings when it is > 0). Moved to 31DEC13 this date. 14348. April 15, 2014 SPFLG Eric The help file was not updated last October when the DPARM(4)=2 option was added. This divides by the flux including a spectral index and is thus channel dependent while =1 uses IF-dependent fluxes only. Moved nowhere. 14349. April 15, 2014 UVFND Eric A call to SOUFIL was put in ahead of UVINIT and so the LUN used by SOUFIL had not been initialized. The call was not needed in fact. Moved the statements that depended on the call to after UVGET which calls SOUFIL. Moved nowhere. 14350. April 15, 2014 SPLAT Eric The header frequency when averaging IFs with BIF greater than one was incorrect by the offset to BIF (added by low level calibration routines and then again by SPLAT). Should go to patch. 14351. April 20, 2014 polarization Eric Testing the (broken) ORI- mode of PCAL/DOPOL has turned up the following corrections: POLSET Needed to pick up the phase of the reference antenna, not the last antenna in the PD file. ORIPOL (a) Pointer used the DTERM for BCHAN rather than each channel in turn. Other similar routines had the channel loop situated so that the pointer was incremented to get each channel, but this routine was different. (b) The values were stored at very bad addresses in the correction array - baselines were separated by 16*NIF words rather than 32*NIF*NCHAN words. They were picked up at the correct stride leading to very bad values applied to the data. LXYPOL Also had the (b) issue above. DATPOOL Dropped warning messages. PCAL (a) In IPCALC (solves the ORI- mode), changed it to handle "no data found" gently, since many channels in a row may be flagged. A message appears and an answer of (0,0) is returned. (b) In LPCALC (solves the simple model), changed it to trap "no data found" and set 0 into the answers rather than going through all the math to end up with 0 anyway. (c) Changed the automatic reading of the previous solution from the AN or PD table into an optional one. The default is to omit that reading - bad values can hurt much more than starting with the assumed circular polarization. (d) Corrected BIF and EIF back to the user's input for writing in the history file. PCAL Help: added CPARM(7) option to read initial guesses from the AN or PD table or, by default, omit that. Moved PCAL changes only to 31DEC13 on 8-Aug (the ORI- mode was found not to work well so no need for such a complicated patch). 14352. April 24, 2014 RLDLY, FRING Eric Both has a memory-saving equivalence that changed the size of a COMMON after the parameter MAXIMG was raised making that size be different in different places in the task. Dropped the equivalence. Moved nowhere. 14353. April 24, 2014 XGAUS Eric Changed XGAUGE (that gets initial guesses) to 1. Require that the average of 3 adjacent points be above the cutoff rather than all 3 points being above. Also allow that test both with and without removing the estimated baseline. 2. There is no initial guess from the user now, so use as one of the guess possibilities the previous solution but with the peaks something like the data values at those pixels. 3. The possible estimate that was the previous solution did not use the previous baseline due to "typo"-level issues. These issue meant that ZEMAN would occasionally find a pixel with no Gaussian solution where it expected one. Moved nowhere. 14354. May 2, 2014 VPLOT Eric The option to plot amplitude and phase together restricted the X axis type to 12 (IAT time). Dropped the restriction. There may be problems when asked to plot models - we'll see. Moved nowhere. 14355. May 5, 2014 IMAGR Eric Add options to plot labels and "stars" on the TV and to default OBOXFILE to make unique ones for each run. Changed: IMAGR Help: add LTYPE adverb, add S commands to BOXFILE, have OBOXFILE = 'ENVariable:' means make a unique name for this run. IMAGR Add new adverb -> CLEAN, parameterize # stars maximum = 100, fill in default OBOXFILE, set graphics channels 5,6,7 for extra labeling, pick up stars from BOXFILE -> CLEAN object STBOXF New routine to parse Stars from BOXFILE. QCLEAN Handle 3 extra graphics plane choices, put them in TV object at each OTVLOD. TVUTIL Add NSTARS, LTYPE as known parameters, in OTVLOD and similar routines pick them up from the parent object and call OTVLAB and OTVSTR as needed. Change OTVLAB to get LTYPE from parent object if TV object does not have it. Add OTVSTR subroutine to pick up parameters and pass to TVDSTR. TVDEVICE New routine TVDSTR to plot those stars that fit on image using graphics routine PNTPLT. BOXFILE Help: add S format cards for stars. Moved nowhere. 14356. May 6, 2014 IMAGR Eric Additional fiddling: QCLEAN Fixed typos and values in default graphics channels TVDEVICE Removed redundant lines, added call to SETLOC in TVDSTR TVUTIL Fixed typo causing failure of compile on SUNS in OTVROF IMAGR Rearranged choice of graphics channels, added messages for OBOXFILE and number of stars read. Moved nowhere. 14357. May 6, 2014 UVPRT Eric Another issue with the printed FREQ. UVGET adds the FOFF(BIF) when it gets the "FREQ" parameter - no need to add it again. Moved nowhere. 14358. May 6, 2014 SPLIT, SPLAT Eric Added defenses for when the user says to average 256 channels and there are only 16 or to do every 256 channels when there are only 16. This led to bad values in the header for frequency, increment, and the like. I guess we assumed users would be careful with these parameters... Moved nowhere. 14359. May 7, 2014 DOFIT Eric Changed FRING, KRING, and CALIB to test if the DOFIT option is used and, if so, to require that REFANT be spcified and that REFANT not be included in the to be fit list. Errors fixed: 1. CALIB tested GSOLV before it was set and so did not tell the user that certain SOLTYPEs are not allowed in the DOFIT option. It did nothing with REFANT and GSOLV. 2. FRING stopped the solution for REFANT when REFANT>0 but did nothing when REFANT was not speified. 3. KRING forbad a solution for REFANT > 0 in all cases, whether DOFIT was used or not. Somehow, this seems suspect. Moved nowhere. 14360. May 9, 2014 DOOSRO Lorant Somewhere on the way the definition of APARM(5) in FRING (# of basebands to solve for) changed from quarters to thirds. Modified the pipeline to correct for this. Note that the explain file for FRING still has the old quarter description for APARM(5)=4, whereas the HELP (top) part of FRING.HLP has the newer definition of thirds. Moved nowhere. 14361. May 14, 2014 CURVALUE Eric AU6B called TBCLOS twice, the second leading to an error message. Moved nowhere. 14362. May 15, 2014 TGET, TGINDEX Eric AU2A was not properly prepared for having no old-format TG files. Fixed the error handling in TGET and set a parameter to avoid an "already closed in FTAB" message. Moved nowhere. 14363. May 15, 2014 TVHUI Eric Help: Clarify OPTYPE, used in all cases Fortran: It did not use the negative options when the transfer function was linear, but it did have different parameters set for slope and offset. Thus, negative linear transfers were a mess in displaying intensity and hue. In writing out the image, only the hue transfer is used, but that too ignored negative when the transfer function was linear. In creating the output wedge file, it used a COPY call when FILL was correct - this led the 5th axis to have 5 pixels on it in my test version but could have done all sorts of things with other compilers. Moved nowhere. 14364. May 20, 2014 FQCENTER Eric Another unintended consequence of the addition of this option is that the source table contains velocities which are the velocities at the reference pixel. This then requires a new subroutine CENTSU to correct the source velocities. It uses the source table routines (SOUINI, TABSOU, and SUINI, SUREFM called by them) and also the FQ routine CHNDAT all of which were in $APLNOT. They have all been moved to $APLSUB with only TABSOU getting some re-typing. Changed: DBCON to warn about re-running SETJY. Remove FQCENTER from CVEL - too complicated given the fact that the source velocity in the SU table is based on the reference frequency. Wrote CENTSU to correct SU table velocities to the new reference pixel. Changed CENTFQ (1) to call CENTSU, (2) to actually write the updated FQ table (it did 2 reads!), and (3) to do the update the harder way by making a new version and then copying back with appropriate deletions. Moved nowhere. 14365. May 20, 2014 TVFLG, SPFLG Eric Some places did not respond to flagged rows in the FC table gently, but instead made it a hard error. UNDO does flag rows, so this is a serious issue. Moved to 31DEC13 29 May. 14366. May 21, 2014 RMSD, CALIB Eric RMSD had to be changed - the shift to a larger MAXIMG made some arrays exceed 2*31-1 bytes. Changed it to limit itself to images no bigger than 32768 on a side. CALIB had some typos that blocked compilation made May 7. Moved nowhere. 14367. May 22, 2014 EDITA, EDITR Eric Changed QEDIUTIL for the usual table reading issue. Error code < 0 => flagged not bad. Moved nowhere. 14368. May 26, 2014 XGAUS Eric 1. Default PIXVAL -> 10^9 rather than 100*FLUX. 2. Fitting error -1 means NITER exceeded. Say so. 3. Increase scratch buffer size to handle larger size characters (needed for memo 118). 4. Allow menu title lines to be 2 lines long (when the TV is small or the characters large). 5. Help file wording improvements, PIXVAL default. 6. The output routines messed up the residual image header when not writing any parameters. This stopped the correct writing of any images. 7. A format was not wide enough when asking to enter one of '1', '2', ... to RETR with that number of Gaussians. Moved nowhere. 14369. May 27, 2014 XGAUS Eric TSKHLP Changed code to substitute the current value of fo the string when displaying to the user. HLPXGAUS Changed this online help to have better grammar using the new capability of TSKHLP. Added a new set of IMAGE definitions when the SHOW is left off to save space. TVMENU Changed the positioning of the menu and spacing of the items. Added test for a menu too large for the display area. XGAUS Changed XGAUTV to pay more attention to the size of the display area. Made the title rows use a tighter format when the title row would be wider than the TV. Changed the handling of columns to go to multiple ones more creatively. This was forced by trying to fit 8 components with a regular size TV and double size characters. Moved nowhere. 14370. May 27, 2014 DBCON Eric One of the two references to the FQCENTER adverb value made 0 true and the other made it false. This was not a good idea and caused it to refuse to function in a simple case. Moved nowhere. 14371. May 28, 2014 DBCON Eric The UV scaling for data set 2 was based on the change in frequency of data set 2. It really should have been based on the change from the frequency reference of data set 2 to the frequency reference of the output data set. Also the logical DOUVM2 should be based on the frequency difference between the initial reference frequency of data set 2 and the final output frequency rather than only the change in frequency of data set 2. Moved nowhere. 14372. May 28, 2014 CLPLT, CAPLT Eric These tasks extended the max/min values by a little to make room to see all the points, but did not do this for the user's DPARM(4->7) or in CLPLT for an extremum selctions of -180 to 180. Changed them to do the extending after application of the adverbs and the extremum, rather than before. This also allows trapping of max = min. The TRIANGLE adverb help file contained obsolete information and was completely re-written. Moved nowhere. 14373. May 28, 2014 XGAUS Eric After a HAND (entire values by hand) operation one may wish to take those values as an initial guess and re-try the fit. Add DO FIT option immediately after a HAND, but not the rest of the time. Adjusted help file for this. Moved nowhere. 14374. May 29, 2014 Slice plotting Eric Found some issues with laebling slices: 1. LABINI - added special code in LTYPEs 4-6,8-10 for slices to label with X or Y plus Z as center when slice is along an axis 2. TVLAB - put the Y axis label string 1 character too far to the left 3. XGAUS - help did not allow LTYPE > 6. 4. XGAUS - add code in scaling routine to correct for actual number of pixels to exceed screen size. 5. TVSLIN - added code to scale plot to fill the screen as much as possible and the code to correct scaling to make sure plot does not exceed the screen size. RMFIT and ZEMAN use the graphics package routines for plotting and so do not need these fixes. Moved nowhere. 14375. May 30, 2014 IMFIT, JMFIT Eric Found more problems with negatives, causing some error bars to be set at some stage to negatives. This gets confused with the flag to avoid fitting those parameters. This led the fit routines to fail to report some valid fit parameters. I think it is all now under control (but I thought that before too). Moved nowhere. 14376. June 2, 2014 HUINT Eric NEW TASK: This task implements the TVHUEINT verb with a TV display and menu approach. It then writes an image in who knows what units with RGB planes and also a corresponding step wedge - all if requested. Put this back although there are display mysteries (in TVHUEINT also) which would be nice to explore (Blue enhancement turns black on the TV not on the output image). Moved nowhere. 14377. June 3, 2014 BLSUM Eric Changed DOCRT so that value 0 means no printer plot. Added OPTYPE='WEIG' with a clip level APARM(3) so that the sum is done weighted by the blotch image values, but only those > APARM(3). Changed code to keep blotch image in memory along with the blotch selection image. Moved nowhere. 14378. June 3, 2014 HUINT, TVHUEINT Eric Changed subroutine HILUT in task HUINT and in the subroutine used by AIPS so the enhancement of the hue image will not make black regions on the screen where deep blue should be. Note that HUINT did not make black regions in the output image, only in the displayed image. Moved nowhere. 14379. June 4, 2014 UVFLG Eric Fixed comment in help about TAFLG. It is OPTYPE not OPCODE. Moved nowhere. 14380. June 4, 2014 IMAGR Eric The graphics overlay channel for DELBOX caused the stars to ne erased. Added code to QCLEAN to cause the image to reload. Moved nowhere. 14381. June 6, 2014 XGAUS Eric Changed it to allow a "hand" setting to fix parameters for the next plot/fit in the current spectrum. This was non-trivial since the code had not had to support fixed, non-zero parameters in a long time. The function routine must be fed in advance the full list of parameter values and must use all the parameters in evaluating the function, not just those being fit at present. Follow up routines must preserve the full list of parameter values, correcting those that were fit but not zeroing those not fit. For fun, also added a plot of the RMS when the model is plotted. This will make it more reasonable to ignore the extensive printouts in the terminal window. Changed XGAUS and HLPXGAUS help files accordingly. Moved nowhere. 14382. June 9, 2014 INDXR, FITLD Eric Added defense in GETMC and GETDEL to avoid a very unlikely zero divide. That code has been there a long time. Moved nowhere. 14383. June 9, 2014 CookBook Eric Changed font in CookBook to palatino except the tt family was kept to computer modern since palatino used a light-weight Courier for fixed spacing fonts. This juggled pagination and recipes some affecting all $DOCTXT/LaCook*.tex. Corrected errors describing RFLAG in the new Low-band chapter LaCookL.tex. One cannot do a tget after running it since that will wipe out the adverbs it just set. Moved the PostScript and pdf files eveywhere. 14384. June 10, 2014 ELINT Eric Changed the limit on the number of OT table entries to 100000 from 10000. Also added test to detect this failure and to view things as an error when TABOTT returns an error. Moved nowhere. 14385. June 10, 2014 IBLED Eric Put in test for menu falling off the bottom of the screen and added meaningful messages about CHARMULT or TV size. Moved nowhere. 14386. June 12, 2014 MODSP Eric NEW TASK: A spectral-line version of MODIM designed to make a spectral-line cube in I and V with Zeeman splitting. It will be need to test ZGAUS when I get to that one. Moved nowhere. 14387. June 12, 2014 XGAUS Eric Yet another silly bit of code meant that the output of images that were not square in the 2 celestial coordinates was likely to fail (if NY < NX). The headers of the displayed images were also compromised. Moved nowhere. 14388. June 13, 2014 MAXFIT Eric Changed MAXFIT help file to add DOINVERS which says to fit for a maximum in the negative of the image - i.e. for a negative going peak. Changed AU9 to fit in image values not absolute value of image values, to use DOINVERS to negate the image values as they are read in, to negate again the peak found, and to label the peak value as a Maximum, Minimum, or Image (for IMVAL verb). Moved nowhere. 14389. June 13, 2014 DTSUM Eric The task omitted the highest numbered antenna from the visibility matrix due to a logic error in deciding which antennas to print. Moved to 31DEC13 this date. 14390. June 16, 2014 Coordinate labels Eric Changed DIRRA and DIRCOS to use a wider search range at all iterations on the new, modern-style whole-sphere coordinate types. Plots starting 1 pixel off the edge gave them grief otherwise. Changed ITICS and CTICS to be more forgiving about pixel coordinates at the edges of plots. "Failures" by 0.01 pixels were too demanding, use 0.1 pixels now. ITICS also used a border pixel number where an edge pixel number was needed (CTICS already had this fix). Moved nowhere. 14391. June 19, 2014 XGAUS Eric Changed the image display to blow up small images by pixel replication. Moved nowhere. 14392. June 19, 2014 RMFIT Eric 1. Fitting error -1 means NITER exceeded. Say so. 2. After a HAND (entire values by hand) operation one may wish to take those values as an initial guess and re-try the fit. Add DO FIT option immediately after a HAND, but not the rest of the time. 3. Add this DO FIT to the help file. 4 A typographical type error meant that images which were not square in the 2 celestial coordinates were likely to fail on output of the residual (if NY < NX). The headers of the displayed images were also compromised. 5 Changed the image display to blow up small images by pixel replication. 6. Made scratch buffers larger to handle larger characters. 7. Corrected an error message 8. Changed to make 3 column edit menu if it will not fit on the TV otherwise. Also made multiple formats for the title lines depending on TV and character size. 9. Changed menu to avoid displaying thickness or spectral index when we did not fit them. 10. Fixed help file explain to have the spectral index and thickness parameters in the menus. 11. Fixed typos in MODIM help file. Moved nowhere. 14393. June 19, 2014 ZEMAN Eric (1) Changed to plot initial guess at data points not on a 1->1000 thing like the Gaussians. It was step-function like for no good reason. Fixed the Gaussians to use 1000 or 1.3*Nchan, whichever is larger. (2) Changed NGAUS to MAXGAU (3) Changed the image display to blow up small images by pixel replication. Moved nowhere. 14394. June 19, 2014 TI2HA, HA2TI Eric Added the REFANT option to both. REFANT=0 means to use the lower numbered antenna in EACH pair to comppute the hour angle while REFANT > 0 means to use REFANT always. This should keep the data more closely time ordered. Moved nowhere. 14395. June 23, 2014 CLIP Eric Added BPARM with (1) > 0 => flag parallel hands when one of cross-hands is flagged. (2) > 0 => flagg all polarizations when any one is flagged. Moved nowhere. 14396. June 23, 2014 IMAGR Eric Changed UVWAIT, MAKMAP, IMGMEM to handle the keywords WTNOISE and SUMWTIN in a consistant fashion - there was confusion between single and double precision. Now all are double. Moved nowhere. 14397. June 24, 2014 MODSP Eric We realized that ZGAUS is not likely to be of use and have abandoned plans for it. Changed MODSP to omit R-L separation but allow separate R and L fluxes, so users can still do what they might want including Zeeman-like things. Moved nowhere. 14398. June 24, 2014 PCAL Eric PCAL used to quit on 10 consecutive failures. Bad channels may exceed this so changed the limit to 256 consecutive failures. Any one success resets the counter. Moved to 31DEC13 on 8-Aug. 14399. June 25, 2014 BLSUM Eric Added the option to weight by the square of the blotch image. Moved nowhere. 14400. June 26 & 27, 2014 BLAPP Eric Added a ton of messages to identify where error conditions are raised, which may help identify what is failing. Also put the SUBARRAY input variable into the UV data object. June 27: added a bunch of input adverbs to the UV data object in case various defaults might not find data. Moved nowhere. 14401. June 27, 2014 CALIB Eric A special routine ECALC is used to solve for gains when there are 2 (or 1) baselines. It did not get the phase only solutions to have correct (1.0) amplitudes. Added code to convert the amplitudes of the gains to 1. Moved to 31DEC13 this date. 14402. June 27, 2014 SNPLT Eric SNPLT did the metric scaling with the actual max and min rather than the intended plot max and min. SNPLT is smart enough to have minimum ranges for all parameters that control the latter. Changed the metric scaling to use the intended plot range. Moved nowhere. 14403. June 27, 2014 FQCENTER, UVCOP Eric The CENTFQ call was placed before the fancy table copying and so did not find the tables it needed to fix. CENTSU had an excess close call. Moved nowhere. 14404. June 30, 2014 WTMOD Eric The u,v,w scaling parameter was only set when FQCENTER was invoked. The u,v,w came out zero when FQCENT <= 0. Moved nowhere. 14405. June 30, 2014 FQCENTER Eric Changed CENTFQ and CENTSU to check that there are FQ and SU tables (resp.) so that their absence is not treated as a bad error. Moved nowhere. 14406. June 30, 2104 AIPSLetter Eric LET14A has been put back in TEX, PS, and PDF forms. Moved widely in $AIPSPUBL. 14407. June 30, 2014 AIPS Memo 118 Eric AIPSMEM118 in TEX, PS, and PDF forms has been writte. It is entitled "Modeling Spectral Cubes in AIPS" and describes in detail XGAUS, ZEMAN, and RMFIT plus all the model creation routines. Moved widely in $AIPSPUBL. 14408. July 1, 2014 FTFLG Eric This task did not get all the fixes given to SPFLG and it should have: 1. The new in-core gridding method left zeros rather than blanks in the times between scans. 2. Gridding counters can overflow (> 2 billion) so it died rather finishing - except that it ignored the error code and got in deep trouble. Counters -> double precision, error codes get checked. 3. If a source number was missing from the source table, the source list was at best incomplete or entirely missing. 4. Changed GRIDFT to have the default source number (INSNUM = the number of a single source) in the call sequence to use in gridding. Changed UVIMG to honor this addition (but with 0). 5. FTFLG labeling worked well when the full image was displayed, but needed a lot of clean up to handle sub-images. 6. Changed ftFLG to have DPARM(4)=2 mean to divide by source flux and do a channel-dependent spectral index adjustment of that flux. 7. Some places did not respond to flagged rows in the FC table gently, but instead made it a hard error. UNDO does flag rows, so this is a serious issue. 8. It was missing the on-line help file entirely and the explain section of the regular help had menu items not in FTFLG (just SPFLG). Moved nowhere. 14409. July 2, 2014 FTFLG, SPFLG Eric Corrected bugs affecting both tasks. The labeling of IF numbers was bad when the sub-image window began in after the first pixel of the first IF. The flagging of data in the display and sometimes even in the main data file and the falg command file failed on sub-image windows and on displays of every 2nd (or more) channel due to an excess of channels times IFs. Moved nowhere. 14410. July 3, 2014 LWPLA Eric Changed the default meaning of DPARM(5) to mean force portrait. Values other than 0, 2, and 3 mean to fit to paper. The help file already had a reference to HELP POSTSCRIPT but it wa given additional info including the strong suggestion that ones uses LWPLA to chose font size and type. The positioning of character strings depends on these and it is a major pain to adjust all the string positions for a font change in the PostScript. Moved nowhere. 14411. July 3, 2014 SNPLT Eric Changed so that it uses the same metric scaling for all plots on a page (based on the first plot on the page) when plotting only one type. It labels each vertical axis when plotting more than one type so they self-scale entirely. Moved nowhere. 14412. July 6, 2014 Plotting Eric Added options to control metric scaling in LTYPE adverb. Changed Fortran: METSCL New routine to adjust metric scaling based on value of LTYPE/100. AU5B Add LTYPE (hundreds part) to TVWLABEL. GAL Call METSCL with 103 IRING Call METSCL with 103 WIPER Add adverb LTYPE, default -7. Handling LTYPE differently AU6E AU9B TKLAB TVUTIL CLAB1 CLAB2 COMLAB IAXIS1 TVLAB SLFIT CCNTR CNTR ISPEC PROFL RSPEC SL2PL XBASL XGAUS XPLOT APCAL BPLOT FGPLT PFPL1 PLCUB PLROW Handling LTYPE differently and calling METSCL rather than METSCA directly: AU5D GTICNT LABINI SLBINI GTIC PLAXES EVAUV BLSUM BPERR GREYS IMEAN IMRMS IMVIM KNTR PCNTR SERCH TAPLT ANBPL CAPLT CLPLT DFTPL FRMAP FRPLT GCPLT PFPL2 PFPL3 PLOTR POSSM RFLAG SNIFS SNPLT TARPL UVHGM UVPLT VPLOT WETHR Help files updated in meaning and allowed range: LTYPE WIPER BLSUM PLAYR SLFIT IMAGR EVAUV BPERR CCNTR CNTR GREYS IMEAN IMRMS IMVIM ISPEC KNTR PCNTR PROFL RSPEC SERCH SL2PL TAPLT XBASL XGAUS XPLOT ANBPL APCAL BPLOT CAPLT CLPLT DFTPL FGPLT FRMAP FRPLT GCPLT OBPLT PFPL1 PFPL2 PFPL3 PLCUB PLOTR PLROW POSSM RFLAG SNIFS SNPLT TARPL UVHGM UVPLT VPLOT WETHR TKGUESS TKMODEL TKRESID TKSLICE TVCUBE TVGUESS TVLABEL TVMODEL TVMOVIE TVRESID TVSLICE TVWLABEL Moved nowhere. 14413. July 9, 2014 VPLOT, ANBPL, CAPLT, CLPLT Eric These 4 tasks could lose track of the user's fixed plot scales and gradually expand them in plot row after plot row. Added an array, set at the beginning, and then used to provide the fixed scales each time they are needed. The adverbs are filled in with the actual values used since they are part of the plot file header and found by PLGET. Moved nowhere. 14414. July 9, 2014 IMAGR, SCMAP, SCIMG Eric TVUTIL used the wrong pointer to call the routines that handle UNClean boxes. Apparently IMAGR had adequate adverbs set to keep errors away. But SCMAP and SCIMG did not. Fixed the calling of the UNClean box display to be when it is requested. Moved nowhere. 14415. July 11, 2014 IMEAN, IMSTAT, TVSTAT Eric Changed IMEAN and AU6D so that units of JY/Bxxxx and JY/Pxxxx would be viewed as able to give total flux in Jyxxxx. Changed IMEAN so that small windows will not change ACTNOISE and ACTMEAN keywords unless DOCAT >= 2. Moved nowhere. 14416. July 11, 2014 BPASS Eric Added 2 new options to normalize with BP function amplitude squared (power) rather than BP amplitude (voltage). BPASSPRM(10) values 5 and 6 are used for ICHANSEL and all channels, respectively. Moved nowhere. 14417. July 11, 2014 PRTSY Eric Major add-ons: 1. In simplistic global averages, added control over what parts are actually printed. 2. Added routine to average each scan with MEDIAN and then to display the Pdif, Psum, and Psys of (pol,IF,ant,scan) with various arrangements: (1) ant by scan per IF, (2) IF by scan per antenna, and (3) IF by antenna per scan. The RMS from the medians is also available for display. Moved nowhere.outprt 14418. July 11, 2014 TVMAXFIT, MAXFIT Eric Fixed wording in MAXFIT inputs about DOINVERS and added that adverb to the procedure TVMAXFIT inputs and help file. Moved nowhere. 14419. July 11, 2014 LWPLA Eric Changed ZLWIO to handle file names longer than 120 characters. Moved nowhere. 14420. July 31, 2014 ANFIND Eric Added array name ATCA to the known arrays, apparently older AN files from the ATCA said something else.... Moved nowhere - relink much. 14421. July 31, 2014 UVPLT Eric Option 17, baseline length projected on a particular position angle, failed since that number was typed as an 18 after the real option 18 was already selected. Moved nowhere. 14422. August 4, 2014 MFPRT Eric Correct the velocity computations. They did not use the correct furmulae and assumed that ALTSW had not been used. Fixed the formulae and checked/handled the ALTSW question. Moved nowhere. 14423. August 4, 2014 VLBI sampling Eric Changed FXSEG to check only the first 4 characters of the CQ table sampling function. HANNING was not always spelled out fully. Moved nowhere. 14424. August 4, 2014 CALIB Eric Changed CLBPA to force solutions of (1,0) with weight SNRMIN+1 on those antennas not being solved (DOFIT option). Otherwise some antennas could end up with weight 0. Moved nowhere. 14425. August 5, 2014 FITLD, FIXTM Eric Added REFDATE to both tasks. In FIXTM, if REFDATE is ahead of the header obs date, then the times will be increased and the header changed to REFDATE (or even earliere if very negative times occurred in the data set). In FITLD, the internal reference date will be changed to REFDATE. If that date is not before or equal the internal reference date in the IDI file, some warning messages will appear and the data will not be loaded. Moved nowhere. 14426. August 7, 2014 GETDATE Eric New verb: GETDATE returns THEDATE in 4 different string formats using the FORMAT adverb. changed POPSDAT, NEWPARMS, AU7B, and help files for THEDATE and GETDATE. Moved nowhere. 14427. August 8, 2014 SAD, IMMOD Eric Upgrade history file: SAD - more digits on number sources line IMMOD - now up to 9 comps, before limit was 4. Add line about INLIST. Moved nowhere. 14428. August 8, 2014 TYSMO Eric Added option to smooth and clip the Pgain's as well. This is needed since there are dropped SY values for some IFs in some correlator configurations, but usually plenty of good data in each scan to correct the missing points. Moved nowhere. 14429. August 8, 2014 PCAL Eric The data averaging over time code was written seemingly by someone (not me!!) who does not understand real data. It assumed that all samples had 4 valid correlations (RR, RL, LR, LL) and that a straight average would be better than a weighted one. If the RL and LR were flagged a lot compared to either one of RR and LL, then the source polarization would appear reduced by this (bad) averaging method, assuming the flagged RL and LR values were not crazy. The weight sums included all weights which could have been negative sometimes or zero, so any weighting done woth those sums would have been questionable at best. Moved to 31DEC13 this date. 14430. August 19, 2014 TVFLG, SPFLG, FTFLG Eric Added the option to flag a range of IFs (when there are > 2 IFs). This changes the FC file format, makes interactive flagging rather trickier, and a variety of other issues. Interactive flagging now draws pink lines only within the confines of the data. Changed the help and on-line help files concerning the SWITCH ALL-IF FLAG option. Moved nowhere. 14431. August 20/21, 2014 TYSMO Eric The flagging code had bad errors - flagging RR on LL flags, not checking for LL good, and messing up the polarization selection of flags (not using the one that applied to the current flag). Added flagging of PostGains (not moved to 31DEC13). Moved bug fixes to 31DEC13 21-Aug. 14432. August 21, 2014 SNPLT Eric Fixed the code to plot XY polarization tables correctly labeled. Moved nowhere. 14433, August 22, 2014 capitalization Eric Changed TKLAB, TVLAB, CLAB1 to capitalize Ascension and CLAB2 to use upper and lower case like the other 3. Moved nowhere. 14434, August 22, 2014 POSTSCRIPT Eric Added to this general help file an attempt to describe how to put two PostScript files together containing the image and step wedge initially produce by HUINT (followed by KNTR and LWPLA). Moved npwhere. 14435. August 22, 2014 SNEDT, EDITA Eric Changed QEDIUTIL to add the Post-detector gains to functions using the SY table. It is plotted and used in editing just like the Pdif and Psum. Only other noticable change was to recompute the Tsys (Psys) after smoothing the Psum and Pdif. Changed help files SNEDT, EDITA, HLPEDISY and HLPEDISS to show the new options. Moved nowhere. 14436, August 25, 2014 POSSM Eric The code that computed the correlation functions, XCF and ACF, ignored the existence of polarizations greater than one. Added a loop over polarizationand polarization in the addressing. Dropped the backwards passage through the loop since the code used an intermediate variable to handle the increase in the size of the data (twice as many points, real only). Doubled the size of the two buffers involved. Moved nowhere. 14437. August 26, 2014 SAD, IMFIT, JMFIT Eric Added adverb EFACTOR to the tasks so that the deconvolution is tried at the 27 possible EFACTOR * (-1. 0. 1) times the uncertainties in the major axis, minor axis, and position angle. Default EFACTOR is 1.3 - previously the code used a hard-wired EFACTOR = 1. Also added code to make a decision about whether the component is probaly resolved, probably unresolved, or somewhere in between and to display this comclusion for the user. Changed wording in the explain sections of the help files about the error computation (giving details of the Condon-based method) and about the deconvolution and the resolved/unresolved issue. Changed POPSDAT.HLP, NEWPARMS.001 for the new adverb and created a new help file. Moved nowhere. 14438, August 26, 2014 CLPLT Eric Corrected error bars. The very old code assumed that the weights were the uncertainties. They are in fact one over the uncertainty squared (when they are at all meaningful). Also added a print (to a file only) option and warnings in the help on DOERRB. Moved nowhere. 14439, August 26, 2014 CAPLT Eric Inherited the error bar code of CLPLT: Changed averaged sigmas to 1/sqrt(weight) and changed the error computation which really was for phase not amplitude. Added DOEBAR warnings about weights being in proper units. Moved nowhere. 14440, August 29, 2014 DRNX.INC Eric Changed the include file to allow 10000 scans (nearly 3 hours of 1-second scans)! For OTF. Moved nowhere. 14441. September 2, 2014 OTF issues Eric Changed LISTR (SCAN option) to allow up to 10000 source names. Changed PUVD.INC to raise XSTBSZ to 10000 from 2500. Means an awful lot of re-compilations and link edits. Moved nowhere. 14442. September 3, 2014 IMFIT, JMFIT Eric Changed IMFIT and JMFIT to use the primary beam corrected total flux and the primary beam and bandwidth smearing corrected peak flux in the "is it resolved" question. Changed the help files (and SAD help file) to point out that the deonvolution is done with the bandwidth-smearing corrected Clean beam and the the relative flux tests are done woth the corrected fluxes (if such corrections are requested). Moved nowhere. 14443. September 3. 2014 VLBI TY tables Eric TABTY Changed so that values in the table between 998.5 and 999.5 are returned as magic blanks. 999 was used as an invalid signal by some VLB systems. QEDIUTIL Changed EDGTTY and EDG1TY to blank values within 0.1 of 999. (Affects SNEDT, EDITA, EDITR) SNPLT In SNPDAT, test Tsys and Tant values for 999 and blank. SNIFS As SNPLT. LISTR In GAINAV, check Tsys and Tant for 999 and maike blank. TYSMO Test on 999 in clip and smooth routines, two places each one. Moved nowhere. 14444. September 3, 2014 Spectral indices Eric Changed: FNDSPX Change call sequence to ask NTERM and return SPIX array (1) Flux at 1 GHz, (2) spectral index, and, if NTERM=2, (3) curvature. SPFLG Change to have DPARM(4) = 3 +> use curvature and change the call sequence of FNDSPXplus add code to apply curvature when fit. FTFLG As SPFLG. UVPLT Change to have DOSCALE = 3 => fit and use the full flux at 1 GHz, spectral index, and curvature, Change the call to FNDSPX and add code to apply curvature when fit successfully. CLIP As UVPLT. UVFND As UVPLT. BPASS Add adverb DOSCALE to control usage of spectral index - < 0 -> none, =2 fit for curvature when fitting, 0 or 1 fit only spectral index if fit on "unknown" sources. CPASS Add adverb DOSCALE to control usage of spectral index - < 0 -> none, =2 fit for curvature when fitting, 0 or 1 fit only spectral index if fit on "unknown" sources. Add call to FNDSPX which was not present in an otherwise good spectral index routine. BLCHN Add adverb DOSCALE as in BPASS to control fitting. PCAL Serious changes: Add DOSCALE to control spectral index usage. Drop simplistic fitting of spectral index. Add subroutine PSOSPX to find spectral parameters of it is a "known" source and subroutine PNDSPX to fit "unknown" sources. (FNDSPX is overkill since frequencies and fluxes are already known in PCAL.) Changed help files for each of these tasks to add adverb or new meanings if the spectral index control was already there. Moved nowhere. 14445. September 5, 2014 BPASS, CPASS Eric Both had a prohibition against linear polarization stating that it assumed that RR/LL was coming from cal routines but DOPOL was false. Since one gets XX/YY from cal routines this is at best silly. Moved nowhere. 14446. September 7, 2014 TYAPL Eric Invent plan to have VLA.EFFICIENCIES file in $AIPSIONS eventually and change TYAPL to read it. By default, it will use the list of frequency-dependent, but antenna and polarization independent values it has always used. GETBND Take subroutine from INDXR and VLAMP, change to handle frequencies in Hz or GHz. Adjust call sequences to cover the full range of usage FNDEFF Take subroutine from TYAPL. Rewritten to use old values only in the absence of a text file. Read user's CALIN if available and looks likely to be correct, else use the standard one in $AIPSIONS: INDXR Give up GETBND routine and adjust call sequences. VLAMP Give up local versions of FNDEFF and GETBND, Adjust call sequences and size of VEFF. TYAPL Add CALIN adverb for efficiencies file. Change the efficiencies values to include polarization and antenna dependence. Call GETBND to find band to use in FNDEFF. Moved nowhere. 14447. September 8, 2014 IMFIT Eric The call to the fitting routine to fit for the image rms was flawed with many parameters not being set to correct values. Eventually found them all. The error code returned is much more complex than was thought as well. Moved nowhere. 14448. September 8, 2014 ACCOR Eric Added FLAGVER adverb and changed the IO to read via UVGET rather than more directly so that the flags can be applied. Moved nowhere. 14449. September 11, 2014 TVFLG, SPFLG Eric Added protection of the current valid header when trying to TV load a new plane. If there was no valid data, then the header was messed up but the program went on happily until it tried to use that header (i.e. flagging the still displayed plane). Moved to 31DEC13 Sept 12. 14450. September 11, 2014 ANTAB Eric Increase local index table arrays. Forbid multiple subarray data - the input text files do not know about that. Moved nowhere. 14451. September 12, 2014 BPASS Eric Increased all dynamic memory allocations by 1 kiloword. They were too tight and the memory pointers could get corrupted. Moved to 31DEC13 Sept 12. 14452. September 15, 2014 MODIM Eric This task overlooked the possibility that real images with an FQID axis would be used! Fixed that oversight. Moved nowhere. 14453. September 15, 2014 BPASS Eric The task correctly identifies linear polarization as Stokes 5 and 6 (in abs value) but then did not handle arrays that go from 1 to 2 only in a proper fashion. Fixed numerous places for this. Moved nowhere. 14454. September 16, 2014 PRTAB Eric Changed handling of defaults for NCOUNT, BDROP, and EDROP. If all 3 are <= 0 or NCOUNT <= 0 and EDROP < BDROP, then all values in a table element will be printed. If EDROP >= BDROP and BDROP > 1, NCOUNT will be set to at least 1 so that things like time, antenna, subarray, etc get displayed. Dropped upper limit on DOFLAG in the inputs and clarified the defaults in the help. Moved nowhere. 14455. September 16, 2014 DOBAND Eric DATBND essentially forbad BP calibration of XX/YY data. Changed one character to treat XX/YY in the same manner as RR/LL. Moved nowhere. 14456. September 16, 2014 SNPLT Eric Added OPTYPEs PDGN and PSGN to plot Pdif and Psum divided by (4096 * Pgain * Pgain). Psum and Pdif are recorded with a scaling by Pgain and this undoes that (relevant when the Post gains are changed from scan to scan). Moved nowhere. 14457. September 17, 2014 XX/YY models Eric Fixed GETCTL so that it can handle an Ipol model with XX/YY uv data. Moved nowhere. 14458. September 17, 2014 SNPLT Eric Fixed the addressing in the new options 'PDGN', 'PSGN'. Moved nowhere. 14459. September 22, 2014 PRTSY Eric Corrected logic handling TIMERANG since records got out of sync. Changed it to print the center of scans rather than some approximation affected by breaks between scans. Corrected handling of scans not selected by SOURCES. Added option to do the median average over source and optionally to divide one source into the others. Moved nowhere. 14460. September 23, 2014 TVFLG, SPFLG Eric The gridding routines in GRIDTB, TVFLG, SPFLG, and FTFLG set up the Stokes of the work file as RR/LL or RL/LR or I but did not honor XX/YY. This made the many preparations to label XX/YY in the rest of the tasks useless. Made the gridder much smarter and dropped some non-functional NStokes=4 things in FTFLG. Moved nowhere. 14461. September 24, 2014 flagging XX/YY Eric The flagging routines used the Stokes values, so XX/YY fell off of the 4-valued Stokes flag array and used the next flag table entry (if any) instead. Fixed DATFLG plus tasks UVCOP, TYAPL, PRTSY, SPLAT, VLAMP, and RFLAG, all of which applied flagging to data and/or tables. Moved nowhere. 14462. September 24, 2014 Stokes strings Eric AXSTRN Add XX, YY, XY, YX FLGSTK Add RRLL, RLLR, XX, YY, XY, YX, XXYY, XYYX, CROS and improve examination of type of data Stokes versus strings. SETSTK Only use is to set the STOKES adverb to select the single polarization (even if input adverb calls for > 1) - added XX/YY. PRTIM Added XX, YY to list UVPRT Adjusts STOKES input and sets info re columns to print - added XX/YY Moved nowhere. 14463. September 26, 2014 CLCOR Eric Added OPCODE = 'POGN' for gain polynomials that are in power rather than voltage units. Added this to help and added emphasis on 'GAIN' that it is voltage units (in help and also history). Moved nowhere. 14464. September 29, 2014 PRTSY Eric Changed the SORC OPCODE to allow normalizing by IF and/or antenna as well as by source. Added an EFFS opcode to do SORC but then read an efficiencies file and write a new one based on the ratio between 2 sources wrt some reference antenna whose effciencies we believe. Moved nowhere. 14465. September 29, 2014 BSCAN Eric Added INVERS. If > 0, then require scans to have at least 1 PC table record. Also added the count of PC table records to break ties between scans (if any). Moved nowhere. 14466. September 30, 2014 MC tables Eric Changed the time forgiveness to plus or minu 5.1 minutes. This should help with the new MC tables that have longer intervals. Moved to 31DEC13 this date. 14467. September 30, 2014 XGAUS, RMFIT, ZEMAN Eric The first 2 of these required the user to supply at least 6 more pixels on the spectral axis than the number of parameters being fit (and RMFIT assumed DOSPIX for all components). Changed this to require only 1 more pixel and to count the parameters correctly in RMFIT. All tasks failed to set error codes when the spectral axis was either too large or too small. This meant that they did not do some essential things but then went on running. Moved nowhere. 14468. October 1, 2014 APCAL Eric Changed APCAL in two ways: 1. The spllover routine has used the VLBA 7mm spillover for everything. Added better spillover models for S band and for the average band (other than 7mm). APARM(3)=2 allows one to have the previous form of the correction if one needs it for continuity. 2. It was noticed that Tsys appears to vary with "IF" by rather more than one would expect. Added an option (APARM(6) > 0), to "normalize" the Tsys's to the average Tsys by averaging over time the ratio of Tsys(IF,t) to the average over IF at that time t. The normalizing factors found are displayed as a function of iF and polarization. Moved nowhere. 14469. October 2, 2014 FTFLG, SPFLG, TVFLG Eric Added the concept of overlapping "pieces" into which the time range is divided when the number of times divided by the smoothing time does not fit on the TV screen. The menus get the option LOAD NEXT PIECE to load the next piece in sequence when the number of pieces is greater than 1. This adjusts the window appropriately and does a LOAD. The number of pieces and their size are reset whenever ENTER SMOOTH TIME is used. Changed Fortran, hep files, and real-time help files. Also corrected the axis labeling code to avoid trying to write characters off of the screen (on OFF WINDOW + LOAD). Moved nowhere. 14470. October 2, 2014 APCAL Eric Added APARM(7) to give IF to use for normalization rather than the average. Also added tests for 999 -> blanked and tests for TS > 0 before using the value. Moved nowhere. 14471. October 6, 2014 Doc files Eric Updated the tab completion file HLPIT.LIS and general ABOUT help files: ADVERB, ANALYSIS, CALIBRAT, EDITING, GENERAL, INTERACT, MODELING, POLARIZA, POPS, SPECTRAL, TASK, TV-APPL, TV, UV, and VERB. Moved nowhere. 14472. October 10, 2014 CookBook Eric Changed chapters: 4 Add TVFLG new menu item (load next piece), IF range flagging, new figure to show new menu, FTFLG mentioned and RFLAG 5 Add mention FTFLG, IMAGR axis labeling, stars, and unique OBOXFILE option. 6 Rewrite hue-intensity paragraphs adding HUINT. Say more about LTYPE. Memo 118. 7 BLSUM WEIG option, MODSP, Memo 118 8 Add HUINT, add ref to AIPS Memo 118 with XGAUS et al. 9 CLCOR 'GAIN', BPASSP(10)->6, ANTAB before USUBA, APCAL new options 13 Update all changed ABOUT files C Change BPASSP(5) from 0 to -1, (10) to 6 in both cases. E Pgain in TYSMO and EDITA and new SNPLT OPCODEs. Spectral index curvature now can be fit automagically. Table of contents and index files 0, I, J also updated. Recipe.tex also rearranged as needed. Moved nowhere. 14473. October 14, 2014 CLIP Eric Corrected the usual omission of XX/YY polarization from the case statement setting up the clip parameters. Also changed ACLIP. Moved nowhere. 14474. October 14, 2014 CLCOR Eric Corrected formats for POGN and GAIN which caused it to die before doing the correction. Moved nowhere. 14475. October 14, 2014 APCAL Eric Rearranged code so that there is no error updating header in catalog when there is an odd number of antennas being plotted. Moved nowhere. 14476. October 16, 2014 UVFLG Eric the shadowing and elevation routines reduced the computation time by a factor of 350 by redoing the needed computations only when the time in the data actually changes. Unfirtunately a change was defined as >= 1 second. With truncations etc data that are 1 second will not always trigger a recomputation. This led the last record in some scans tp be omitted from the flags. Changed the time extension from 100 to 90 millisec and made the time changed test the same. Moved nowhere. 14477. October 16, 2014 APCAL Eric Changed code so that it continues computing if the plot option is turned off (DOTV 1 initially, button D hit on page full). Previously it just closed down. Moved nowhere. 14478. October 16, 2014 VLBARUN Amy VLBA pipeline. The major changes from VLBAPIPE are: -- removed all reference to tapes -- added ionosphere and EOP corrections -- added many sensible defaults (CLINT, REFANT, SOLINT INTERPOL, TIMERANGE for instr phase can scans etc.) -- put in levels of plotting -- allow plots to be written to disk, converted to .gif and put in an html file for easy viewing -- autoboxing in IMAGR -- optional e-mail message stating when pipeline is done Moved nowhere. 14479. October 17, 2014 SPFLG Eric The fast-gridding path through the code did not implement the spectral curvature and did bad things when that mode was requested. Moved nowhere. 14480. October 17, 2014 KNTR Eric Panel labels LABEL=2 did not give the metric scaling. Changed it to use AXSTRN with the relative coordinate value. Changed AXSTRN to use more adequate formats and kHz as well as Hz and MHz. Moved nowhere. 14481. October 17, 2014 LISTR Eric Found numerous uses of ICOR0 that did not include XX/YY polarizations, leaving a variety of curious labeling of our new P band stuff. Moved nowhere. 14482. October 20, 2014 POSSM Eric Fixed Stokes labeling for X/Y in the bandpass and other similar subroutines. (UV data were already correctly labeled.) Moved nowhere. 14483. October 21, 2014 ACSCL Eric NEW TASK: ACSCL is a version of ACCOR but with many more adverbs. ACCOR adjusts the amplitude scale so that autocorrelations (prior to APCAL) remain at 1.0 despite issues with bandpass normalization and the like. In fact, ACCOR may be omitted in the processing and replaced, after FRING and BPASS, with ACSCL. Note that ACSCL requires DOCAL and DOBAND >= 1. Added comment about ACSCL in BPASS help at BPASSP(10) and BPASSP(4). Brought BPASSPRM help file up to date. Moved nowhere. 14484. October 22, 2014 Misc. Eric Changed: 1. GCPLT help file inputs section not correct on use 2. IMFIT add SNR display on peak intensity or corrected one if the peak has to differentiate between header and channel clean beams 3. JMFIT as IMFIT. Moved nowhere. 14485. October 23, 2014 KRING Eric The logic associated with adverb DOFIT was defective causing it to fail. A missing CALL to MSGWRT made that failure even more mysterious. Moved nowhere. 14486. October 23, 2014 FRING et al Eric Include file GAIN.INC parameterized its arrays as 2*MAXANT but at least FRING used them up to 3 * max antenna number. So long as antenna numbers are less than about 60 this should still work (I hope). LOFAR data has antenna numbers higher than 60 and FRING failed. Moved nowhere. 14487. October 24, 2014 MFITSET Eric NEW VERB: Changed AU5, POPSDAT, NEWPARMS and write new help file for a verb to set essential adverbs for JMFIT and IMFIT from the TV (name adverbs, BLC, TRC, NGAUS, initial guesses). Moved nowhere. 14488. October 31, 2014 TVFLG, SPFLG Eric Added LOAD LAST PIECE option as well to Fortran, help, and run-time help for TVFLG, SPFLG, and FTFLG. Moved nowhere. 14489. November 5, 2014 MFITSET Eric Changed AU5 and the help file to add the option to plot the Gaussians as they are set. Moved nowhere. 14490. November 5, 2014 TVSAD Eric NEW TASK: SAD with TV interactivity to adjust initial guesses, island windows, and judge whether one believes the answer. Adjusted the TITLE and menu relative position a little in TVMENU. Moved nowhere. 14491. November 5, 2014 SAD Eric Adjusted some formats and made a fix to avoid RMS computation failure due to a central, one box, spike in the histogram. Moved nowhere. 14492. November 25, 2014 INP, INPUTS Eric Added adverb DOPRINT which, if <= -2, will suppress the page full stops in interactive mode. Batch and RUN modes already ignore page full. Changed AU1A to implemet this and the 2 help files. Moved nowhere. 14493. November 25, 2014 SNPLT Eric Changed the SUM plotting to use the user's requested SYMBOL. Moved nowhere. 14494. November 26, 2014 TVSAD Eric Corrected: 1. 2 strings that were too short 2. Plot routine for Gaussians and set Gaussian position angle by 90 degrees due to internal form of the position angle in most of SAD (not final output). 3. Subtract the Gaussian over a wider range of rows in making the resdiual. 4. Added the Gaussian number in the center of each Gaussian on the cyan graphics overlay 5. Added the option(s) to REDO individual Gaussians, leaving the rest unchanged. Moved nowhere. 14495. November 26, 2014 ABOUT lists Eric Updated the TAB complletion file HLPIT.LIS. Updated the ABOUT help files for ANALYSIS, CALIBRAT, INTERACT, MODELING, POPS, RUN, TASK, TV, UTILITY, UV, VERB, and VLBI. Moved nowhere. 14496. November 27, 2014 CookBook Eric/Amy Changed chapters 6 and 7 to add MFITSET and TVSAD, chapters 9 and C to add VLBARUN, plus ABOUT lists, table of contents and index. Moved nowhere. 14497. November 27, 2014 TVSAD Eric Added display of ellipses for the Gaussians as one is setting them (in graphics plane 5). Once set, 5 is cleared and the ellipses are plotted in plane 4. Involved making an ellipse drawing subroutine to do the work from wherever it is needed. Moved nowhere. *********************************************************************** *********************************************************************** Changes while 31DEC14 is NEW *********************************************************************** *********************************************************************** 14499. December 4, 2014 SAD, TVSAD Eric Changed: SAD: Subtract Gaussian over wider area, TVSAD suggested that the previous setting really did not go far enough. TVSAD: 1. Add CURVALUE option to both menus. 2. Have CURVALUE run when setting the Gaussian parameters. 3. Allow up to 8 (parameterized) Gaussians in any one island (only under user command). Also TVSAD and HLPTVSAD help files edited. Moved to 31DEC14 this date. 14500. December 6, 2014 TVSAD Eric Corrected the menu placement which should not expand if the character size expands (and it was too far in in any case). Moved to 31DEC14 this date. 14501. December 11, 2014 Misc Eric Changed 1. KWIKHD to understand XX/YY polarization 2. SAD to subtract components over a much wider range of rows 3. TVSAD as SAD Moved to 31DEC14 this date. 14502. December 11, 2014 AIPS Memo 119 Eric AIPS Memo 119, describing the use of the new TVSAD task, has been put in $AIPSPUBL in PS (very large), PS.gz, and PDF forms. $AIPSPUBL is in all releases. 14503. December 11, 2014 UVSUB Eric removed simple prohibition to allow XX/YY to be modeled with I models. Moved to 31DEC14 this date. 14504. December 11, 2014 CENTFQ Eric Changed this subroutine, called by numerous tasks, to adjust the FQ table for centering the frequency reference channel. The routine did not consider that the column numbers and pointers might differ from the input to the output files. But if an old FQ with no bandcode comes in a new one with blank bandcode will go out. So sepearte numbers/pointers for the two files are required. Moved to 31DEC14 this date. 14506. December 12, 2014 CLIP Eric The type of clip level to use for V was compromised by a typo. Should work now. Moved to 31DEC14 this date. 14507. December 15, 2014 HUINT Eric Add remarks about HUINT to help files for TVHUI and TVHUEINT. Also corrected typo in AIPS Memo 119. Moved to 31DEC14 this date. 14508. December 15, 2014 INPUTS Eric Setting the page full suppression at DOPRINT=-2 hits too many other uses - so changed it to <= -9. Changed AU1A and help files for INPUTS and INP. Moved to 31DEC14 this date. 14509. December 18, 2014 CORFQ Eric Changed it so that FACTOR < 0 causes the uvw to be scaled but not the header and antenna-table header frequencies. This can be used to undo the unfortunate re-scaling of holography "u,v,w"'s in almost all AIPS tasks. Moved to 31DEC14 this date. 14510. December 18, 2014 UVHOL Eric Added adverb DPARM to specify the max weight, amplitude, and u,v for the display. The self-scale routine remains present if DPARM(1) < 0 but is skipped unless requested. Corrected the display of channel and frequency to use the correct average channel and frequency computed with the average channel and BIF. Moved to 31DEC14 this date. 14513. December 30, 2014 SNPLT Eric Test on outside plot range not on correct plot types (tested REAL) and type 24, not the PC phase. Moved to 31DEC14 this date. *********************************************************************** *********************************************************************** Changes while 31DEC14 is NEW & FROZEN *********************************************************************** *********************************************************************** 14516. January 2, 2015 DOFARS Eric Changed the help file to emphasize things about BLC and TRC. Values 1,2,4,5,6,7 are applied in TRANS, BLC(3), TRC(3) are applied in the FARS step. Dropped OUT2NAME et al from inputs and help and DOFARS.001 - they were left over from earlier versions of FARS. Moved to 31DEC14 this date. 14517. January 8, 2015 DOFARS Eric I messed up the proc (close quotes issue). Moved to 31DEC14 this date. 14518. January 13, 2015 RMFIT Eric Task needs to copy FQ table (if any) to output residual images but not the model fit images. Moved to 31DEC14 this date. 14519. January 13, 2015 VLBAUTIL/VLBATECR Amy When VLBATECR calculated which files it needed to download it did not handle observations that went over the new year correctly. Fixed. Moved to 31DEC14 Jan 19. 14522. January 16, 2015 VLBARUN Amy Fixed two small gotchas in VLBARUN. If a user starts aips with "notv" then then default TVCHAN=1 is invalid, this caused KNTR to die which then caused VLBARUN to die; changed TVCHAN=0. Second if the command "convert" does not exist on system and OUTFILE is set (i.e., the user wants output plots in a html file) VLBARUN would run, but no plots would be produced and the html file would be useless. Changed VLBARUN to check if OUTFILE was set that "convert" exists, and not to run if it does not. Also undated HLP file to say that "convert" is required for those features. Moved to 31DEC14 Jan 19. 14523. January 16, 2014 FTFLG Eric Some vestiges of the way SPFLG handles baselines remained in the code and they were not correct for FTFLG. It led the output flag table to be incorrect when one or all baselines to one antenna were specified in the ANTENNAS and BASELINE adverb. Moved to 31DEC14 this date. 14525. January 20, 2015 BPASS Eric One array was dimensioned 500 and should have been MXBASE. ALMA with 36 antennas breaks this badly. The history routine listing antennas also had a typo bug and was not general and so failed in this case. Moved to 31DEC14 this date. 14533. January 29, 2015 MORIF Eric There was a bug in the handling of BP-like tables that made a mess of the results. Bug was in REIFS but did not hurt there. Moved bug fix to 31DEC14 this date. 14540. February 11, 2015 TYSMO Eric the clip parameters (APARMs) were not applied unless the MWF cutoff parameters (CPARM(6-9)) were set. Corrected this to apply the APARMs with or without the CPARMs. Moved to 31DEC14 this date. 14546. February 18, 2015 XAS Eric Character writing into grey-scale memories used a background of 1 which means that the characters could disappear with some LUTs. Changed it to zero in XAS and in $YGEN/YCHRW. Moved to 31DEC14 20 May. 14550. February 21, 2015 bandpass Eric The low level, VLB routine DETRAT was still in left-hand antenna coordinates. This was known to the bandpass application routine BPASET but not elsewhere, particularly BPASS. Therefore BPASS made solutions which were not good when applied to the calibration data at least for high-frequency, narrow-channel data. Changed DETRAT to be right handed and BPASET to go along with this. Changed some vestigial bits in BPASS that had no effect. Moved to 31DEC14 this date. 14581. April 23, 2015 UVINIT (!!!!) Eric UV data sets are always supposed to be larger by 4 kilobytes than they actually have to be. But we have encountered some that are exactly the right size for some reason. In that case, if the data set is small enough to fit in the first IO buffer, there should be no read into the second buffer. The loop thatstarted the first 1-2 reads skipped the second if the count was < 0, but my reading of the code said that the count would never be < 0 for the first buffer but could be zero. Changed the LT to a LE. Moved to patch and 31DEC13 (for Huib). 14578. April 21, 2015 START_AIPS, TVSERVERS Josh Malone/Eric Changed ^/tmp to ^/\(private/\)*tmp in both scripts to account for the new usage in Mac OS 10.10.3 (Yosemite). With the correction Mac yosemite is friendly again to simple aips commands with no TV= options. Moved to 31DEC14 24 April. 14598. May 19, 2015 Cal routines Eric GAININ Computed wavelength as if the reference pixel in frequency is always 1. It is now usually Nchan/2+1 making problems with dispersion and polarization calibrations. DATCAL Required both antennas to have non-zero dispersion to apply the correction. This leaves out all baselines to the reference antenna. Moved to 31DEC14 this date. 14600. May 20, 2015 XAS Eric Test for Mac's odd DISPLAY variables dir /tmp but not /private/tmp which is now used. Added the latter so that shared memory will be used on Macs. Moved to 31DEC14 this date. 14604. June 17, 2015 KRING Eric This version of KRING does not work all that well. But it cannot work when the help file has fewer adverbs than the Fortran. Put the 31DEC15 help file back to correct this; the 31DEC15 changes to the Fortran are to extensive for a patch. Moved from 31DEC15 this date. 14607. June 22, 2015 IMAGR Eric The code that did the baseline-length time averaging copied too few spectral channels worth of data to the output buffer in the final dump of all remaining baselines (by the number of random parameters). Moved to 31DEC14 this date. 14660. August 14, 2015 PCAL Eric When doing models (at least), the task failed to notice when a channel was fully flagged and died later. Changed so that the initial data copying would notice this problem. I suspect that the route through the task for solving for a model noticed the absence of data later and handled it. Also fixed unused labels and, with global equivalences, some mismatches. Moved to patch this date. 14669. August 22, 2015 SETJY Eric One routine had 3 different references to format 1000. Merged two of them and made a separate format for the other. Raised MAXSOU to 5000. Moved to patch this date. 14682. September 3, 2015 SAD, TVSAD Eric The creation of the ST table caused the new MF table to disappear from the header. Changed to use the correct header version in the STINI call. Moved to patch 7 Sept. 14687. September 14, 2015 PCAL Eric The SPECTRAL false mode was broken - it did not handle the end-of-file as a normal confition but as an error. Moved correction only to 31DEC14 this date.