[daip] mysterious problem with fittp

[Eric Greisen]

Diana Worrall writes:
 > 
 > I'm running into an fittp problem when using the 31DEC04 version of
 > aips.  Both DS9 and kvis (from the KARMA package)
 > think the sky coordinates of my output FITS file are the same but
 > very wrong values (~0 hr, ~1 deg instead of the correct value
 > of ~16 hr, ~17deg).
 > 
 > When I run fittp from the 31DEC03 version of aips, the sky coordinates
 > are correct are interpreted correctly by kvis and DS9.
 > 
 > I don't see an obvious difference in the KEYWORD values in the
 > headers of the old and new files, but there must be something.
 > 
I can't be much help without more information - you haven't told me
even what kind of data you have - image, single-source UV,
multi-source UV.  What would be the most help would be the PRTTP
outputs with PRTLEV=1 or even 2 for the 31DEC03 and 31DEC04 disk
files.  A possibility is that karma reads all header words including
history files and there is something in the history that is confusing
karma.  But that is moderately far fetched.  The only changes in FITTP
from 31DEC03 to the present are to use D format when appropriate in
format statements and to insure that the table buffers are 0 filled.

retrieving revision 3.12
retrieving revision 3.9
diff -r3.12 -r3.9
36c36
< C;  Copyright (C) 1995-2000, 2002-2004
---
> C;  Copyright (C) 1995-2000, 2002-2003
255c255
<          FDVEC(1) = 129 - NTAPE
---
>          FDVEC(1) = 30 + NTAPE
535c535
<                WRITE (CARD,1011) HKEY, KEYTS(IKEY), KEYWOR(IKEY),
---
>                WRITE (CARD,1010) HKEY, KEYTS(IKEY), KEYWOR(IKEY),
578,579c578
<  1010 FORMAT (A,I1,2X,A8,' =',1PD19.11,' /',A)
<  1011 FORMAT (A,I1,2X,A8,' =',1PE17.9,' /',A)
---
>  1010 FORMAT (A,I1,2X,A8,' =',1PE17.9,' /',A)
1516c1515
<  1220 FORMAT (A8,'= ',1PD20.11,' /',A)
---
>  1220 FORMAT (A8,'= ',1PE20.11,' /',A)
1634c1633
<  1030 FORMAT ('HISTORY AIPS WTSCAL = ', 1PD18.11, 1X,
---
>  1030 FORMAT ('HISTORY AIPS WTSCAL = ', 1PE18.11, 1X,
3163d3161
<       CALL RFILL (720, 0.0, TAPBUF(TBIND))

 The only changes since the 02 version are to add a loop over catalog
number.

Looking in the subroutine and include areas I also see nothing that
could cause this.

Please send me the PRTTP outputs.

What does aips think the coordinates are after reading back these
files?

ERic Greisen