[SpectralDM] - UTypes changes

CresitelloDittmar, Mark mdittmar at cfa.harvard.edu
Sun May 4 20:26:00 PDT 2014


All,

A quick note about ObsCore and it's relation to the 'Cube' model diagrams.
ObsCore 'defines the core components of the Observation data model', and
there is no Observation data model.

The Cube project 'Dataset Model' diagram is the result of evaluating the
ObsCore, Spectral, Char models and folding in the discussions from the dm
list (Observation-Dataset).  If you compare ObsCore Appendix B with this
diagram, you will find a near exact match of attributes.  The differences
are:
  Target.Class => Target.objectClass (because 'class' is a poor name for an
attribute when implementing)
   Access.* => Not included, this is a DAL extension to the
ObservationDataset and would be owned by ObsCore
   Provenance => Node is gone.. the contents are referenced directly from
the dataset.  This was the result of dm list discussion.

All Char domain utypes (if they match Char1.13) are consistent because I
just reference the Char model.  As pointed out, this comparison has been
done by more than one of us, and is in pretty good agreement.

So, this diagram becomes the "Observation data model" (Observation Dataset
model actually) of which ObsCore is showing a 'core component' view.

====
Second item... I would be fine delaying Spectral2.0 to finish folding it
into this effort.  I have already done the UML interpretation (in Volute),
but have not done a good review and I'm sure there are mods to be made.  My
point for this thread, was just that this latest draft was not intended to
go that far.

====
> I think it requires however constructive adjustments with respect to what
already exists.
I'm looking forward to hearing what you have in mind!

Mark
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.ivoa.net/pipermail/dm/attachments/20140504/0c07e538/attachment.html>


More information about the dm mailing list