SED data model v0.92

[Doug Tody]

Hi David -

>From David Berry:
> Am I correcting in thinking that if the spectral coordinate
> is linear with channel number, you still have to supply all the spectral
> coordinate values explicitly, rather than just giving a scale and offset?

Yes, this is the case at present.   The main reason we did this was
to avoid getting into all the issues involved in specifying a general
spectral WCS.  A fully populated spectral coordinate vector is dead simple
but handles all the various cases (simple linear, irregularly sampled, log
lambda, polynomial fit, spline fit, etc.).  Plus the same approach works
consistently for the time axis and so forth.  Much spectral data in archives
is in this format, and probably all 1D spectral data can be represented in
this way.

It might be worthwhile to add support for a simple linear dispersion
(scale and offset) as a special case.  For the cost of slightly more
interface complexity we would gain some efficiency of representation.
But there is a good chance this would escalate and lead to reinventing
spectral WCS.  It might be better to keep it simple now, and just add
general support for FITS WCS later as we expand the VO data models and
XML representations to cover this area.

Note the FITS representation already supports the proposed FITS spectral
WCS, using the -TAB convention to map the WCS spectral coordinate onto
a field of the binary table.  Although we don't include real WCS support
in SSA currently, it should be easy to use the coordinate vector to build
a WCS on top of SSA data.

	- Doug