spectral WCS in FITS

[Robert Hanisch]

As many of you know, substantial progress has been made in the last year or
two on standardizing the representation of world coordinate systems in FITS.
Two papers have been published thus far, describing both a general formalism
and specific implementation for celestial coordinates.  A third paper is in
near-final form, describing spectral coordinate systems.  I was asked to
read and comment on this paper, and would like to draw wider attention to it
and raise a concern I have to see if it is shared by others in the VO
development community.

You may find the latest draft of this paper here:

    http://www.aoc.nrao.edu/~egreisen/scs_112103.ps.gz
    http://www.aoc.nrao.edu/~egreisen/scs_112103.ps


My concern is that the paper provides no simple FITS ASCII table
representation for spectra in which the wavelength scale is given by an
explicit list of values, each with an associated flux.  Rather, it supports
a wavelength scale look-up using a FITS BINTABLE extension, and only allows
the wavelength and flux arrays to be stored as vectors in single cells.
There are advantages to this approach in terms of efficiency and robustness.
However, looking ahead to the imminent development of a Simple Spectral
Access Protocol, I would expect that a basic encoding scheme for spectra
would be a VOTable with columns for wavelength, flux, error, etc.  It seems
to me that this VOTable should be as close in structure as a FITS
representation as possible, to the point where we should be able to
translate the VOTable to FITS with an XML style sheet.  Also, there is much
extant software that reads and writes spectra in even simpler ASCII text
files, with tab- or space-delimited columns and very simple column headers.
Thus, in the interest of the simplest possible interoperability for spectral
data, I have suggested to the authors of the FITS spectral WCS paper to
consider adding a FITS ASCII table option for the tabular form of a
spectrum.  So far I've been told that there is "no way in which it can be
implemented."  I do not believe this statement, but would like to hear from
others in the VO community before pressing the case further.

Thanks,
Bob