Adding New Keywords to enable Conditional Requirements.

[Andrew Robbertz]

There are some keywords which are required for all FITS files. However, there are are also some keywords which are conditionally required based on the file contents and the instrument being used.

For example, depending on the type of the observatory, different keywords are used to accurately detail the location of the observatory at the time the measurements in the FITS file were taken.

• If the observatory is ground-based then the OBSGEO-X/Y/Z keywords are required.
• If the observatory is earth-orbiting then the GEOX/Y/Z_OBS keywords are required.
• If the observatory is deep-space then the DSUN_OBS and HGLN/LT_OBS keywords are required.

Similarly, depending on the king of instrument, different keywords are required to describe the instrument mode and configuration used for taking the measurements.

• If the instrument is a spectrograph, then many additional keywords are required, including SPECSYS, VELOSYS, WAVEMAX/MIN/REF/UNIT.

From reading the recommendations, it seems like there are not any keywords that can be used to determine which if any of these conditional requirement should be applied.

• It does not seem like there is a keyword that can determine if the observatory is one of ground-based, earth-orbiting, or deep-space
• It does not seem like there is a keyword that can determine if the instrument is an imager or a spectrograph

To enable these conditional attribute requirements for the SOLARNET metadata standards, I'd like to introduce two new keywords:

• OBSRVTYP: Type of observatory. A string (from a limited/discrete list) uniquely identifying the type of observatory.
  • This details whether the observatory is ground-based, earth-orbiting or deep-space and can be used to determine the required metadata attributes for the observatory.
• INSTRTYP: Type of instrument. A string (from a limited/discrete list) uniquely identifying the type of instrument.
  • This details whether the instrument is Imager or Spectrograph and can be used to determine the required metadata attributes for the instrument.

[Stein Haugan]

Hmm.... good point(s).

I was going to argue that they have little practical use, but then I realized that they are important from a validator's point of view...

BTW the HGLN/-LT/DSUN_OBS keywords should never trigger a complaint - a calculation of what has been observed will always go through those coordinates (or equivalents), so having them there never hurts. The OBSGEO vs GEOX keywords should of course be mutually exclusive, though.

For INSTRTYP, there are quite a few flavors that should be included: imagers, spectrographs, scanning (or rastering) spectrographs, then polarimeters, then spectropolarimeters, scanning spectropolarimeters. And the new-fangled optical fiber-based multi-point variants (which would require tabulated spatial coordinates btw).