Astronomical Image Processing System
31DEC04 Version


The 31DEC04 version of AIPS has been frozen for about one year except for a few major bug fixes. It is no longer available for installation or for the MNJ. You should install the soon-to-be frozen version 31DEC05 or the current development version 31DEC06.


The final report on 31DEC04 is now available as 31DEC04 AIPSLetter dated 31 December 2004 (PS, 908 Kbytes) has information on all changes in this release. It is also available gzipped, (400 Kbytes) and PDF, (203 Kbytes). The 6-month progress report on 31DEC04 AIPSLetter dated 30 June 2004 (PS, 900 Kbytes) has information on changes before about that date. It is also available gzipped, (400 Kbytes) and pdf format, (200 Kbytes). The list below is a partial summary updated 29 December, 2004.

  1. DELZN has additional output options.
  2. Interactive box setting has improved messages.
  3. UVPLT self-scales correctly on binned data.
  4. CLIPM can now clip on data weights.
  5. UV CC file subtraction and division shows better progress messages.
  6. SNPLT/WETHR do fixed-scale on the X axis when appropriate.
  7. DBCON does proper concatenation of IM, MC, PC, AT, CT, OB, GC files - important for VLBI.
  8. DSORC is a new task to renumber sources.
  9. FITLD messed up FQ IDs when a file had unused FQs.
  10. Changed handling of tape LUNs to allow up to 35 tape devices in AIPS.
  11. Handling of the weather table was corrected in UVCOP, FLGIT, SPLAT, and UJOIN.
  12. CALIB, FRING, KRING can now advance the solution time by sub-intervals of SOLINT.
  13. OHGEO can now handle transposed images.
  14. IMFIT and JMFIT now return their results and offer more control over the rms used to estimate the errors.
  15. TGINDEX and SGINDEX offer display ordering options (date, name).
  16. AIPS was revised to be able to use a read-only file system having AIPS data already loaded to it.
  17. SETFC returns NFIELD, CELLSIZE, IMSIZE adverbs and allows the user to control how much error is okay at the maximum W point
  18. Improved testing on the just-read LASTEXIT file to detect damaged ones
  19. VLBAARCH is a new Run/procedure to convert VLBA data for archiving.
  20. CALIB now uses the V polarization in the source table, allowing WSRT total-intensity calibration mostly
  21. RLDIF does the R-L phase difference for polarization calibration
  22. FILLM tests antenna files to avoid concatenating data with incompatible antenna lists and otherwise to handle data from the archive more smoothly
  23. XHELP now works again and does minimum-match and full cross-referencing no matter how many symbols are in a line
  24. The CookBook is now available in fully cross-linked pdf and html formats
  25. SYSTEM is a new verb to run a command at the level of the shell
  26. Calibrator models are now provided with AIPS using new task CALRD to read them and new verb CALDIR to list them
  27. The divide by channel 0 option now works in POSSM
  28. Errors in handling phase wraps in SNFLG were corrected
  29. POSTSCRIPT is a new help file with useful info and tools
  30. SELF mode in CLCAL was made to do interpolation and to loop over sources, the old nearest-neighbor version is called SELN now
  31. The MNJ will now run POPSGN, build XAS, and update $AIPS_ROOT procedures automagically.
  32. CONVL will accept images that are blanked
  33. MFPRT will print in STARS format
  34. ATMCA is a new task to calibrate astrometric data having multiple calibration sources in better ways.
  35. LISTR does a propoer phase RMS again, but correctly this time
  36. RANDOM is a new verb to generate a random number, GRANDOM generates a semi-Gaussian distributed random number.
  37. Reversed spectra in ISPEC and POSSM were labeled incorrectly
  38. Line printing has new options to limit pagination and repeated page headers.
  39. VPLOT now plots ratios of coprrelators as advertised
  40. PCNTR now allows any of the 4 images to be the contour and/or the grey-scale image
  41. PCAL no longer allows the no-reference antenna method
  42. FINDR is a new task to find normal values of visibility data parameters
  43. FLAGR is a new task to flag visibility data that deviate from normalcy
  44. The WENSS catalog provided by AIPS was corrected to eliminate the sources that are the sum of individual components also in the list
  45. TVFLG and SPFLG now does vector amplitude rms and rms/average as well as scalar. SPFLG does LAST BASELINE as well as NEXT BASELINE.
  46. The automatic starting of AIPS TVs has had several changes which should improve the $DISPLAY setting and/or use
  47. DRAWBOX is a new verb to draw Clean boxes on the TV
  48. DDT and Y2K have new master images/data sets and Y2K has a HUGE test now
  49. EHEX and REHEX verbs now use/return adverbs
  50. Model computations from small or non-power of 2 images now are forced to use DFT; gridded methods are too inacurate
  51. Weights in gain and fringe solutions are now more correct by default and may be modified to make things less weight-dependent (some data converge better with weight=1/sigma rather than 1/sigma**2)
  52. The switch to SDI Cleaning has been changed to a more sensible parameterization.
  53. DFQID is a new task to renumber Frequency IDs
  54. UVFND now does the VCLP option as intended. It was a mess.
  55. FIXBX is a new task to convert Clean boxes from one BOXFILE's cell size and faceting to those of another.
  56. GAL does much larger images, plots correctly with user controls, and returns the fit in adverbs.
  57. CLIPM will clip on weights that are too large as well as too small.
  58. SMOOTH may be applied before or (now) after banpass calibration is applied.
  59. Clean beam plotting was added to GREYS and now has the option of blanking the regular plot in the beam plot area.
  60. CALIB and bPASS offer robust gain solution methods as well as single iterations methods. CALIB reports and can flag on the closure errors from this.
  61. EDITA and EDITR were changed to flag in ways that allow more correct UNDO and REDO options.
  62. A binary installation and MNJ is now available.
  63. and much more....

The final report on 31DEC03 is now available as 31DEC03 AIPSLetter dated 31 December 2003 (PS, 3272 Kbytes) has information on all changes in this release. It is also available gzipped, (1310 Kbytes) and PDF, (1251 Kbytes). The progress report 31DEC03 AIPSLetter dated 30 June 2003 (PS, 750 Kbytes) has information on changes before about that date. It is also available gzipped, (320 Kbytes).

The final report on 31DEC02 is in the AIPSLetter dated 31 December 2002 (PS, 2906 Kbytes); also available gzipped, 1140 Kbytes). See also 31DEC02 AIPSLetter dated 30 June 2002 (PS, 753 Kbytes) for information on changes before about that date. It is also available gzipped, 320 Kbytes).

Distribution Statistics

A tool to monitor shipments of the tar ball was finally installed in Spring 2003. We shipped the 31DEC03 tar ball to 416 "sites" (unique IP addresses) in the last 30 weeks of this calendar year. During the full calendar year, over 551 sites contacted our cvs server. A plot of sites that have had at least one cvs access and of sites that have copied at least one of the tar balls is generated each week. The cumulative totals are quite significant. The cvs access implies that the site installed the 31DEC03 version of AIPS (MAKE.MNJ is not run while installing the frozen version) and/or ran the MNJ for 31DEC03. The tar-ball downloads count only downloads that completed successfully. In each of the totals, an individual IP address is counted only once. In general, an IP address is used by only one user, but more than one IP address will be used by those users that use dial-up connections. Thus, these totals are a modest over-estimate. Note, the totals also include multiple computers from single institutions such as NRAO.

Since 31DEC04 was created and 31DEC03 was made slushy and then frozen, we have kept separate statistics. They are plotted again by cumulative distinct IP addresses. From December 2003 into December 2004, a total of 196 different IP addresses downloaded the frozen form of 31DEC03 and 808 IP addresses downloaded one or more copies of the 31DEC04 tarball. Fully 797 IP addresses accessed the NRAO cvs master. Each of these has at least installed 31DEC04 and 231 appear to have run the MNJ on 31DEC04 at least occasionally. The total number of unique IP addresses in these three lists was 1276.

Since 31DEC05 was created and 31DEC04 was made slushy and then frozen, we are keeping separate statistics. Binary installations and MNJs are included in the version statistics and also reported separately. They are plotted again by cumulative distinct IP addresses. The numbers are suprisingly large actually, but they report, nonetheless, only transactions done after 4 December 2004. By the last week in November 2005, 242 sites had downloaded copies of the frozen version of 31DEC04, 48 of these used the binary installation/MNJ option.

Modified on $Date: 2012/01/03 16:03:24 $
[Eric W. Greisen]