[SpectralDM] - UTypes changes

[Douglas Tody]

So for example, in the latest (and earlier) SDM draft we have:

     Char.SpatialAxis.Coverage.Location.Value

Whereas in ObsCore and the ImageDM (and Char2) as of last fall:

     Char.SpatialAxis.Coverage.Location.Coord.Position2D.Value2.C1
     Char.SpatialAxis.Coverage.Location.Coord.Position2D.Value2.C2

Others differ as well, although I have not done a careful inventory.

I have not followed all the discussions, but as this has nothing to do
with Dataset, and is inconsistent with ObsCore/ObsTAP, I suspect there
is an inconsistency.

 	- Doug


On Fri, 2 May 2014, CresitelloDittmar, Mark wrote:

> Doug,
>
> The utype list of the latest draft should be pretty similar to the previous
> version.
> The Change Log is fairly detailed about what was modified.
>
> All changes were in response to the request for these models to share a
> core data model and are founded on the DM list discussion from Nov->Jan
> on the separation of Observation-Dataset (see ImageDM Twiki).
>
> The Cube model work has gone a step further, to separate Dataset from
> ObsDataset,
> but is completely compatible with what is in Spectral.  My expectation is
> that
> with the completion of the Cube model work, a definitive
> "Dataset/ObsDataset Metadata" model
> would be established (either in that doc, or as a separate one) for use in
> Cube/Spectral/ObsCore etc.  The Spectral doc would be updated to remove that
> portion from its content, and reference this other.
>
> Mark
> =====================================================
>
> In Response to:
> The same point should be made about the SpectralDM 2.0.  I just checked,
> and the Utypes in the latest draft do not agree with ObsCore or Char.
>
> There was a time once, about 3 years ago, when we were trying to fast
> track an update of the SpectrumDM to a generalized SpectralDM and wanted
> compatibility with the existing standards and implementations at the
> time.  However if are still now finalizing SpectralDM 2.0, and have
> decided that ImageDM should extend ObsCore, I suggest that the new
> SpectralDM should do so as well.
>
> All of these primary data models - ObsCore/Char/ObsTAP, ImageDM, and
> SpectralDM, should at this point share the same core data model based
> upon the latest Char2 Utypes.  We already pretty much reached that point
> last fall.
>