[fitsbits] FIT Keyword data dictionaries

[Mark Calabretta]

On Mon 2000/11/06 08:48:08 -0000, Patrick Wallace wrote
in a message to: Mark Calabretta <Mark.Calabretta at atnf.CSIRO.AU>
and copied to: Doug Mink <dmink at cfa.harvard.edu>, fitsbits at nrao.edu

>the equatorial, galactic, ecliptic systems are always right-handed.  But
>we're stuck with a left-handed convention for hour angle (west +ve),
>though perhaps that's of limited relevance for FITS.
>
>The real bugbear is azimuth, where a left-handed convention (north thru
>east) is firmly established.

Quite true.  WCS paper 2 says to encode it as a negative azimuth (and use
elevation rather than zenith distance) though it also points out the
possibility of flipping projection plane coordinates via the CD matrix and
using positive azimuth.  However, it doesn't specify coordinate system
codes for either (they would have to be distinguishable) nor for (HA,dec)
since maps aren't drawn in these coordinates.

Anyway the majority of spherical systems in use are right-handed, and
throwing in another left-handed system would introduce another source of
confusion - and we already have enough of those.  So if there's a choice,
as claimed for terrestrial longitude, it's better to be consistent and
stick to right-handed systems.

Mark Calabretta
ATNF