[Cube/vo-dml] ivoa datatypes

[Carlos Rodrigo]

Hi Mark

Thanks for your answer. It's most useful.

Actually some days ago I did a prototype implementation of the spectral2.0 datamodel for the case of
a SSA for theoretical spectra and it reminded this doubt to me.

The three basic SSA operations for a Kurucz theoretical spectra are implemented as, for instance:

metadata:
http://svo2.cab.inta-csic.es/theory/newov/ssap2.php?model=Kurucz&format=metadata

a data query:
http://svo2.cab.inta-csic.es/theory/newov/ssap2.php?model=Kurucz&teff_min=4000&teff_max=5000

a particular spectrum:
http://svo2.cab.inta-csic.es/theory/newov/ssap2.php?model=Kurucz&fid=23

If you ever have time to take a look to it, I would appreciate if you could comment on how adequate
the serializations are.

If you find the time, maybe you can answer just to me (for not spamming the list) unless you think
that the conversation is of general interest.

Thanks

Carlos

On 06/05/14 22:12, CresitelloDittmar, Mark wrote:
> 
> All,
> 
> I'll add my $.02 from the spectral model side.
> Your example snippets point out that the same metadata of name, unit, ucd.. is used in 2 places in
> the model
>    + once for the Data axis.. often serialized in the FIELD element
>    + once for the Char axis, which summarizes the content of a Data axis.
>         here, they indicate which data axis is being characterized, and/or holding defaults for the
> data values.
> 
> I agree that the serialization of this information is not clear, especially for VOTable, and is a
> bit of a pain to
> implement.  The Spectral model Section 3.9.1 is supposed to clarify how the Char and Data axis info
> relate.   
> 
> I think the hierarchy is supposed to be
>   Char.*Axis.Name == matches CharAxis with DataAxis
>   Char.*Axis.UCD == UCD for CharAxis and default for the DataAxis  values
>   Char.*Axis.Unit  == Unit for CharAxis and default for DataAxis values.
> 
> So, unfortunately, I think you would want to serialize the Char PARAMs, and if you then also include
> them (ucd,unit) in the FIELD, they would be explicitly providing those values for the Data points.
> 
> Hope that helps..
> 
> Mark
> 
>