Roy Williams wrote:
> Writing this question is my action from the NVO TechWG phone call this
> week.
>
> There was a long discussion of VOTable and its use in SOAP services and
> other sophisticated XML applications. The complaint is that some XML
> tools do not work properly with the VOTable schema because the SCHEMA IS
> RECURSIVE. People mentioned Axis version 1.3 wsdl2java, and some
> code-binding tools. The recursion is in two places: (a) RESOURCE
> containing RESOURCE elements and (b) OPTION containing OPTION elements.
> The VOTable schema is legal. But the tools clearly do not work with all
> legal schema. We thought back in 2002 that the tools would "catch up"
> and the problem would evaporate. But now it is four years later. Is it
> time for a pragmatic simplification?
>
> QUESTION for this group:
> Is it desirable to build an official standard "VOTableLight" schema
> which does not have recursion, yet validates most instances of VOTable?
> -- The simplified structure would have RESOURCE contains TABLEs, rather
> than RESOURCE contains TABLEs and RESOURCEs.
> -- It could handle a list of tables, but not a tree of tables.
> -- Every "light" table would validate under the full schema.
>
> Roy
>
>
> California Institute of Technology
> 626 395 3670
Hi all,
If we decide to define an alternative VOTableLight schema, concretely, it means that this light schema will certainly become the new VOTable standard (all VOTable files will have to point to this new schema to be correctly ingested by axis 1.3/SOAP library).
For our part, we already use some VOTables with RESOURCE containing RESOURCES, for instance to describe instrument apertures in VOTable (VO-tech footprint action ESO/CDS and STSCI/CDS collaboration for the HST apertures).
I do not say that I vote against this proposition... but just to say that a such modification will be not just a pragmatic simplication if it implies in a medium term to modify the VOTables using this recursion capability.
Regards
Pierre
Received on 2006-04-24Z11:14:32