Reports from data providers

LEDASMetaData kindly provided by DuncanLawGreen

I interviewed Jim Lewis and Peter Bunclark about DMs w.r.t VISTA: InitialVistaDmNotes.

-- GuyRixon - 29 Apr 2004

I have drafted a data model for the INT wide-field survey (WFS): IntWfsDataModel.

- GuyRixon - 04 May 2004

MERLINMetaData (developmental version) -- AnitaRichards - 05 May 2004

Observation data model revision

Here is a preliminary ObservationDataModelRevision for comment and criticism. AnitaRichards - 05 May 2004

Activities

Review existing relevant DMs against general usability and AstroGrid priorities. Focus on aspects needed for use by VO tools currently in use or being built to access priority data e.g. data which is already being used, which is vital for science cases or is from data centres which are keen to get involved - the 80:20 principle.

1. IVOA documents and discussion
   
   IVOA DM page incl. Quantity V0.2.ps or Quantity V0.2.pdf and Observation V0.2.ps or Observation V0.2.pdf
   Debates at IVOA DM forum espec. SIA evolution proposal and heirachical models i.e. IDHA
   Also see IVOA Data Model FAQ

   Discussion point: I do not agree that data models will make UCDs obsolete; e.g. in the above FAQ "We anticipate that we will use the UCD semantic vocabulary to tag data model objects with meanings." DMs should stop us trying to overload the function of UCDs, but I can't see the original use - to go from the obscure names astronomers use for columns and date definitions to terms which VO services understand - being superceded.

2. AstroGrid data providers
   Please can people post up data models o.n.o. for VISTA, CASU, etc. NB a list of metadata is not a data model but a basis... we have to see how much we can fit the necessary parts into existing models as the first step.
   
3. Use of data models
   We need to decide what the priorities are for AstroGrid use and what is needed to be added to DMs to implement these. How do DMs connect to the query language and registry schema? We also need to see if the needs drive us to a position on IVOA debates e.g. a preference for heirachical or flat models.

4. Structure, versioning and provision for updates
   1 high level data model, many low level models. Map between low level models via high level. Are we agreed that we need a workable top-level workable within a few months? If there is one 'top-level' basic model at one time, can any required ammendments be covered in lower-level models/translation layers until a major revision (at >>less frequently than yearly intervals?)

Flat and/or Heriachical: a suggested view

-- AnitaRichards - 22 Apr 2004

Condensed -- AnitaRichards - 07 May 2004