[Obscore Update v1.1] an attempt to describe the Redshift Axis

[Douglas Tody]

On Fri, 7 Mar 2014, Louys Mireille wrote:

> This raises the question of other axis reference positions , which were
> not included explicitely  in the previous Obscore recommendation, but
> defined implicitely by default ( ICRS, TT, TOPOCENTER) I suppose. Please
> correct me if I am wrong.

The ObsCore model as used in ObsTAP fixes all the units and frames to
make simple ADQL queries possible.  If we are extending ObsCore/ObsTAP
to support redshift then we should consider if it makes sense to make a
similar restriction.  I agree that in the full data model, used for
something other than ObsTAP, we may want to permit coordinate frames and
units to vary, as is the case in ImageDM and SpectralDM for example.
But if ImageDM extends ObsCore (ObsTAP in practice) then the issue will
come up.

Note also that there is some support for derived redshift values (also
variable amplitude etc.) in the Derived class.  We should be clear
regarding how this differs from a "redshift axis".  Does redshift really
require an axis?

 	- Doug