proposed VODataService changes

[Patrick Dowler]

On 2008-5-20 04:49, Ray Plante wrote:
>       <table>
>          <schema>...</schema>
>          <name>...</name>
>          <description>...</description>
>          <column>...</column>
>          <column>...</column>
>          ...
>       </table>

So this would say that the schema name is an additional thing one needs to 
know (along with the table name) to write the correct unambiguous table name 
in a query. From a usage point of view, I don't see a big problem. 

However, a schema itself would have content (description, reference to some 
data model whose instances are stored there, maybe a URI/URL to documentation 
if that is normally allowed outside the description). Would one repeat that? 
Declare the schema once at top level and then refer to it?

Schema is very much a namespace concept and things (tables and functions) 
belong to a namespace (maybe the default namespace). Given the shallowness of 
the hierarchy, it doesn't seem worth mapping it into being an attribute. In 
practical usage, some people use the catalog as the namespace and always use 
the default schema, while others use one catalog and multiple schemata. I 
think some RDBMSs actually enforce one or the other through implementation 
limitations. 

-- 

Patrick Dowler
Tel/Tél: (250) 363-6914                  | fax/télécopieur: (250) 363-0045
Canadian Astronomy Data Centre   | Centre canadien de donnees astronomiques
National Research Council Canada | Conseil national de recherches Canada
Government of Canada                  | Gouvernement du Canada
5071 West Saanich Road               | 5071, chemin West Saanich
Victoria, BC                                  | Victoria (C.-B.)