Internal header 
  
  The header is composite of 3 arrays each storing different 
  data type.  The arrays are :
   dbl_keys r*8
   chr_keys character*80
   int_keys int*4 
  These arrays are in common and can be access vis a routine.
  There are currently 4 headers 2 for the integer map 
                                2 for the exposure map
 
  There are 3 access routines to the header one for each variable type
   
  ghead[dis] (mapid, key , val, mode, status)   
 
  mapid : string to ID which map is in used 
  (MAP, MAPSV, EXMAP EXMAPSV)
  key   : string which ID a particular value in the header 
  val   : value of the key 
  mode  : 0=read/get 1=write/set

  NOTE Routines that deal only with the header do not need to have 
       pass down from ximage any argument regarding the header
       Internally the routine can call the function isloaded("map")
       to test if there any values set in the header and return if false. 



 Double array  
 crval1    value in TCRVALm or CRVALm (typically ra for --TAN)     
 crval2    value in TCRVALn or CRVALn (typicaly dec for --TAN)
           Note These value are keept as read from the file 
                (unless precession occurs)  
           These are degrees 
 crota2    value in TCROTA ot TROTA (valid for --TAN)
           Degrees
 ximnorth  north old internal ximage convention 
 cdelt1    value in CDELTm or TCDELm 
 cdelt2    value in CDELTn or TCDELn
           Note ximage now stores the values as read from the file 
                no need to store them negative for non interger 
                increment value. Stored as degrees BUT check 
 ***            internally when used as arcsec.

 ddelt1    value in the DDELT keywords   
 ddelt2  
 drpix1    value in the DRPIX  
 drpix2 
           Note this is an internal ximage / xselect convention 
                and no widely used 
           These value and the C or T value are all kept 
           and available to use in the code
   
 crpix1    value from CRPIXm or TCRPIXm  
 crpix2    value from CRPIXn or TCRPIXn
           Note ximage now stores them as read from the image 
                no need to shifting around the coordinates to get the 
                half pixel

 equinox   equinox. This is the image equinox namely the crval values are
           in this equinox and thier values are always kept as in the 
           original file.
           However :
           The ximage convention of the equinox is the following 
           There is a default equinox set, if an input image has a 
           different coordinate equinox from the default 
           the coordinates are always precessed to the default equinox
           when their are used. Ex 'header' the coordinates are    
           display in the default equinox. 

           If in the reader the qualifier ra and dec are specified 
           these are interpreted as given in the current ximage equinox
           read/ra/dec         equinox=default ximage  
           If instead in the reader the equinox is given together ra and dec
           the internal ximage equinox is changed to be the input equinox.
           read/ra/dec/equinox  the default equinox is changed to the input 
                                equinox
           The following commands 

           cey 2000
           read/ra2000/dec2000  
              are equivalent to 
           read/ra2000/dec2000/equinox=2000  

           default ximage equinox is defined and stored in the main program
                                  in a variable named 'equinox'  
  
 exposure  exposure kept in seconds. For events files is calculated 
           internally based on the gti. For images read from file 
           the following keywords are tried 
           EXPOSURE ONTIME XS-ONTI LIVTIME TIME_LIV XS-LIVTI TIME_NET
           EXP_TIME TELAPSE. 
 ***       Check internally for minites transformation

 dtime     dead time kept as fraction (0-1) read from DEADC DTIME 
           DEADC DTCOR XS-DTCOR DEADTIME

 datamin   image min calculated internally always but never scaled 
 datamax   image max calculated internally always 
           Note different from how was before where scaling was calculated 
           if over 32000. Now the scaling is still needed because internally 
           maps are stored in intergers.

 The following values are mainly for recording, and are not used in
  calculations, etc.

 ocrpix1   Reference pixel for axis 1 in original image
 ocrpix2   Reference pixel for axis 2 in original image
 ocrval1   Reference value for axis 1 in original image
 ocrval2   Reference value for axis 2 in original image
 ximscale  Scaling factor applied to original image
 ximzero   Offset XIMAGE has applied to original image
 

 Character array 

 file      input file name 
 root      input file root (file - extension)
 object    from OBJECT keyword
 ctype1    from CTYPE or TTYPE (if --TAN  coordinates)
 ctype2    from CTYPE or TTYPE (if --TAN  coordinates)
 cunit1    from CUNIT
 cunit2    from CUNIT
 telescop  from TELESCOP 
 instrume  from INSTRUME 
 detnam    from DETNAM
 filter    from FILTER
 date-obs  from DATE-OBS
 date-end  from DATE-END

 Integer array 

 loaded   0 = not loaded 1= loaded 
 expmap   0 = exposure map, 1 = not exposure map
 vignapp  0 = vign correction not applied, 1 = applied
 szx      x size in pixel. This is always the array size read in
          NOTE that if the image has black spaces in it
                these are a kept  
 szy      y size in pixel
 zmx      x zoom. This is an input paramater but can be different 
          from the input beacuse the mission databases record 
          the original detector size. when read a known mission
          the pixel size is compare to the mission databases 
          pixel size and their ratio give the zoom       
 zmy      yzoom. currently ximage do not use the zooms for the two axises
 emin     energy minimun 
 emax     energy maximum