[daip] Re: FITS-IDI and AIPS problems again

[Chris Flatters]

LLoyd,

I'll take these issues in a somewhat different order than they were raised.

1 - Coordinate systems and IMAGR

    I don't understand what you mean by the output map being labelled as
    being in NCP L,M coordinates. If IMAGR is given data in NCP coordinates
    it will produce a map in RA---NCP and DEC--NCP. This appears to be what
    you have. AIPS will label right ascension and declination correctly
    and IMPOS, for example, will return the correct position for a point
    in the image.
    
2 - Epoch zero in source table

    This turns out to be due to VLBA FITS-IDI files not quite conforming to
    the standard. After the next midnight job FITLD will handle both standard
    SOURCE tables and the anomalous VLBA SOURCE table.
    
3 - Equinox not propagating to images

    The equinox of the image is taken from the appropriate source table entry
    so you were seeing a symptom of the source table epoch being trashed.
    
4 - Differences between FITS-IDI headers and AIPS headers

    There is not a one-to-one correspondance between AIPS headers and FITS-IDI
    headers. Some FITS-IDI elements are discarded others are mapped to other
    AIPS elements. For example a FITS-IDI file might have both an OBSCODE and
    an OBSERVER but an AIPS header only has an OBSERVER slot: FITLD puts the
    OBSCODE in the observer slot because it is the more important of the two.
    
5 - FREQID is -1 in antenna table

    This is correct. The FREQID in the AIPS antenna table is used for book-
    keeping related to polarization calibration. You should not try and force
    it to another value.
    
6 - IATUTC = 0.0 in antenna table

    This is also correct. AIPS is only interested in the offset between the
    time system used for the data and UTC (DATUTC). In olden times AIPS data
    was stored in IAT and IATUTC was used instead of DATUTC. The IATUTC
    keyword is retained for compatibility with older AIPS applications but
    should contain the same value as DATUTC. FITLD has correctly identified
    that you data is in UTC and set IATUTC = DATUTC = 0.0 ignoring IATUTC
    in the FITS file.
    
    I notice that your antenna tables have UT1UTC = 0.0 which is unlikely
    to be correct. This would cause some damage if you tried to run UVFIX.
    
7 - FITS-IDI support in AIPS++

    The Jodrell Bank people have been tasked with providing support for
    FITS-IDI in AIPS++. I would estimate that if everything went well
    and the AIPS++ people made a determined effort to do this then it
    would take at least a year (6 months to write the code and another
    6 fiddling to get things right). In actuality, it will take much
    longer.

	Chris Flatters
	cflatter at nrao.edu