[fitsbits] DATE-OBS: circulate the WCS papers!

[Doug Mink]

Steve Allen wrote:
> 
> On Tue 2002-01-22T14:32:12 -0800, Steve Allen hath writ:
> > In effect, we deem DATE-OBS to be broken/deprecated and aren't
> > spending any engineering time fixing how we handle it.

> As a result we feel that we cannot re-engineer newer systems to assign
> a precise meaning to DATE-OBS.  If we did it could be mistaken to
> imply that the precise meaning applies to older instrument systems.
> We feel that we must use privately-defined observation timing keywords
> until such time as a WCS paper V or VI comes along to deal with
> "Representations of temporal coordinates in FITS".

As someone who writes a widely-distributed analysis package, RVSAO, which
utilizes both time and pointing direction keywords, in computing radial
velocities from spectra, I hope that temporal coordinate agreements come
sooner rather than later.  For my work, I need the midtime of the observation,
which is usually computable from start or end times and exposure.  Sometimes
an observation may have temporal gaps in it, making such a computed midtime
wrong.  Temporal coordinates quickly get complicated enough to deserve as
much of a paper-length treatment as spectral coordinates are getting.

I got so frustrated at the myriad combinations of keywords for timing,
observing location, and pointing direction that I wrote an IRAF task to
compute the velocity correction I needed accepting input in as many ways
as I could think of.  Here is a web page describing what it looks for:

 http://tdc-www.harvard.edu/iraf/rvsao/bcvcorr/bcvcorr.proc.html

I also wrote a date conversion program as part of the WCSTools package
which is described at

 http://tdc-www.harvard.edu/software/wcstools/getdate.html

-Doug Mink
 Telescope Data Center
 Smithsonian Astrophysical Observatory
 Cambridge, Massachusetts