[daip] Re: FITLD not conform AIPS Memo No. 102

[Eric Greisen]

Harro Verkouter writes:
 > 
 > Hi Chris,
 > 

      Chris no longer works for NRAO.

 > If I'm not mistaken, the behaviour of the FITLD program, when loading
 > IDI-files, is not according to the description in AIPS Memo #102.

      The software goes to great lengths to try to support any column
order.  FITLD is so complex that I would not be suprised if there is
some gotcha.

 > 
 > In all the table-column definitions it is explicitly mentioned that 'The
 > order of the columns does not matter.'. It is my finding that order
 > *does* matter. At least in one case: the 'SOURCE' table. Accidentally
 > (and unknowingly) I swapped two columns of the 'SOURCE' table. When
 > trying to load the resulting FITS file into AIPS it crashed. When I
 > swapped the columns back to their original position, the file loaded
 > just fine.
 > 

       Would it be possible for you to prepare a modest data set with
a reversed column order?  Then I could walk through with the debugger
and try to find out why the elaborate indexing is failing.

Thanks, Eric