Data Models in Kyoto

[Doug Tody]

Hi Guys -

As Alberto says, Characterization is a key data model for DAL since we
need it to characterize datasets in a standard way.  Much of the SSA data
model deals with generic characterization/identification etc. issues which
are not specific to spectral data.  We will need much this same metadata
for the next version of SIA.

I think the sessions are actually 3-3.5 hours.  It might be possible
to address both characterization and SSA in the joint DAL/DM session.
Or, we could make a start on Characterization in the Tuesday session and
continue with the details in the DM session on Thursday.

BTW, in thinking about the issue of generic dataset metadata it can
be useful to look at SSA.  While the SSA data model needs more work,
it demonstrates more or less what is needed without being overly complex.

	- Doug


On Fri, 15 Apr 2005, Alberto Micol wrote:

> Dear Jonathan,
> 
> >
> > We have three DM sessions scheduled
> > of two hours each and I propose
> >  Tues  pm  DAL/DM joint session (SSAP/SED)
> >  Thurs pm  Characterization and related issues     Clashes with 
> > Registry, Grid-Web
> >  Fri   am  STC, and Catalog/Quantity               Clashes with UCD, 
> > Grid-Web
> 
> 
> Since I'm really eager to see the next generation of DAL protocols 
> (SxAP v2?) based on a symmetric data model, I think it would be worth 
> discussing Characterisation
> *before* we meet Doug et al. in the joint DM/DAL session.
> 
> I know that the main topic for the joint DM/DAL is SED/SSAP, but we 
> also have
> to plan for the future, and seeing whether we are at level of
> maturity that could allow us to move to a Characterisation-based DAL 
> protocol
> in the next year should be discussed before we merge with DAL.
> 
> Tough to accomodate in the current scheduling though; could we have the
> joint DM/DAL on Thursday afternoon?
> But there might be other dependencies I'm not aware of...
> 
> > I'm hoping we can move the Fri session to Wed. so I can attend the UCD
> > session.
> Yes, that would be very nice indeed.
> 
> Alberto
> 
>