Program:      WFITS

Purpose:      Writes disk set and subsets onto magnetic tape (or disk)
              in FITS format.

Category:     FITS, TAPES, UTILITY

File:         wfits.c

Author:       K.G. Begeman

Keywords:

   INSET=     Set (and subset(s)) where to copy data from.
              Maximum number of subsets is 2048.

   BOX=       Frame of input subsets [entire subset].

   FITSFILE=  Name of file to write FITS data to [write to tape].
              If more than one subset is to be written to FITS file,
              a generic filename should be entered. A generic filename
              contains fields of consecutive hash (#) symbols.
              The filenames of the FITS files will be generated by
              replacing each field with the subset sequence number.
              The number of hash symbols in a field denotes the field width.
              Examples:
              FITSFILE=FILE#.FTS     generates FILE1.FTS,   FILE2.FTS   ...
              FITSFILE=FILE##.FTS    generates FILE01.FTS,  FILE02.FTS  ...
              FITSFILE=FILE###d.FTS  generates FILE001.FTS, FILE002.FTS ...
              If no hash symbols are present in a generic name, one hash
              symbol will be appended.
              If data is to be written to file, the following two keywords
              will be skipped.

   OUTTAPE=   Name of output tape device [list of all tape devices].

   SKIP=      Number of files to skip on OUTTAPE [SKIP to End of Tape].
              The first input file will be stored at this location, the
              second at the next location etc.
              Note that for new (empty) tapes you should give SKIP=0.
                                          
   BITPIX=    Specify number of bits to represent pixel value [-32].
              This number can be 8, 16, 32, or -32 for IEEE floating point.
              For antenna patterns it is advised to take BITPIX=32 (or -32).

** BLOCKED=   Specify the number of FITS records / tape record [10].
              The maximum blocking factor is 20.

** MNMX=      Always calculate min, max Y/[N]?
              If the minimum and maximum of the subsets are not found in
              the header, they are automatically calculated. If you are
              not sure that these header values are correct, give MNMX=Y.

Description:  FITS: A Flexible Image Transport System, is a format for
              the interchange of astronomical images and other digital
              arrays on magnetic tape. This format provides a simple but
              powerful mechanism for the unambiguous transmission of
              n-dimensional, regularly spaced data arrays. It also
              provides a method for the transmission of a virtually
              unlimited number of auxiliary parameters that may be
              associated with the image. The parameters are written in a
              form which is easily interpreted by both humans and
              computers (D. C. Wells, Astron. Astrophys. Suppl. Ser. 44,
              (1981) 363-370 ).
              WFITS creates a FITS header record from the GIPSY header
              items and converts the image data (INSET=) to 8, 16, 32
              bits integers or to 32 bits IEEE floating point (BITPIX=).
              Programs to read FITS files like RFITS, use this header
              item to convert tape data to image data.

Updates:      Oct 21, 1991: KGB, Modified code for new mtopen routine.
              Nov 12, 1991: VOG, 'LONG_MIN' casted twice before assigning
                                 to float because of bug in 'gcc'
              Dec  9, 1991: VOG, Put OBJECT in fits header on users
                                 request.
              May 14, 1992: VOG, New defaults for MINMAX=
              Jul 20, 1993: KGB, Completely rewritten, replaced fts_ and
                                 ftsd_ routines.
              Apr 20, 1994: KGB, Minor repairs.
              Feb  7, 1996: KGB, Keyword FITSFILE= implemented.
              Okt  8, 1999: VOG, Removed non printable characters in record
                                 in gdsd_rfits.
              Feb 15, 2010: JPT, Write IEEE NaN for blanks when BITPIX=-32.
              Feb 25, 2013: VOG, Default BITPIX set to -32