[daip] aips related

[Eric Greisen]

Dharam Vir Lal writes:

Please use daip at nrao.edu even though I answer most queries.

 > I am doing simulations (generate visibility data sets) for SKADS using the 
 > simulator in aips++, where i am dealing with LARGE number of antennas.  
 > Subsequently, I use classic AIPS to understand the generated UVDATA.
 > 
 > Since, the number of antennas are more than 50, I see that the AIPS task 
 > UVCOP would not allow me to make a copy of some part of the parent UVDATA 
 > file.

     Certain arrays of POPS adverbs are limited to 50, AIPS internally
understands up to 90 without recompilation.  I doubt sincerely that
you need more than 50 in the arrays ANTENNAS and BASELINE.

 > 
 > Therefore, I am planning to perform a new installation by changing the 
 > hard coded number 50 (= max number of antennas) to a larger number, e.g., 
 > in file UVCOP.FOR from line 560 to 598 in the file UVCOP.FOR (stating 
 > "Find number of selected antennas".  Will this solution work?

   NO.  The adverb in POPSDAT.HLP will have to be changed and then
EVERY task that uses that adverb (proc ADVERB <adverb> will tell you
all that do) will have to be changed.

 > 
 > Also, in AIPS task LISTR lines 2512 onwards and once again lines 3930 
 > onwards, number 28 (VLA data with 28 antennas) seems to have been hard 
 > coded.  Although the AN table is present with each UVDATA file, does this 
 > mean that here as well i will have to change it to a higher number?

     LISTR is not confined to 28 antennas although it may have been
optimized for 28.
 > 
 > Do you know if there are other key AIPS tasks, where I might have to make 
 > such changes?

The main file that needs to be changed is $INC/PUVD.INC which defines
MAXANT.  Note that the standard FITS interpretation of the random
parameter BASELINE is 256*A1 + A2 which limits the number of antennas
to 255.  Miriad has a trick by which a larger number are allowed.  BUt
the number os kep in floating point unfortunately so there are accuracy
limitations to worry about too.

Eric Greisen