VOTable discussions

[awicenec at eso.org]

Hi Francois,

no you did not understand anything wrong, but as far as I could see up to now
the VOTable 1.1 document does not contain the schema anymore and I guess the
reason is that there are actually problems creating that schema using a DTD to
Schema converter. Since VOTable is currently based on DTD it is using some
specialities which you would not do designing the same thing using a tool like
XMLSpy. If you try to use the DTD generated VOTable schema in conjunction with
Castor or Jaxb you will actually stumble over the same kind of problems we had.
Try to model VOTable with UML and you will see the mixture between concepts very
clearly. Yes, I have used the VOTable schema in my example, but just because
that's what we need to have as input to our framework, but that does not improve
the quality of the schema as a data model for tables. I'm actually in favor of a
framework which would keep the current VOTable as one possible representation,
but at the same time I would like to discuss about a possible generalization in
order to be able to describe the same and also more complex data structures in a
more model oriented manner, which would at the same time be more compliant to
the XML world. If we agree to keep the current VOTable as a representation most
of the arguments we've seen in this discussion suddenly are very weak. The XML
framework has the big advantage that you actually are able to convert fairly
easily from one XML format to another and even to something which is not XML
anymore. We should probably make use of this flexibility. 

Cheers and good night from Chile,
Andreas

Quoting Francois Ochsenbein <francois at vizir.u-strasbg.fr>:

> 
> Well, before throwing away all VOTable definitions and specs,
> please keep in mind that the VOTable definitions were set up for 
> INTEROPERABILITY: VOtools can gather data from different horizons, 
> and generate meaningful results from this diversity. Creating a new 
> XSD for each new dataset -- even if each schema is actually derived 
> from some common root schema -- will inevitably generate a wide
> diversity and complexity which has a fair chance to fail and to
> discourage the newcomers... 
> 
> As it was pointed out in the avalanche of comments, the VOTable
> conventions have introduced a possiblity of comparing the METADATA
> of tables generated in different contexts: before being able to combine
> any set of numbers extracted from different data sets, it seems essential 
> to me to compare their metadata -- which physical meaning, which unit, 
> which scale, etc. Only once all these details are known, the comparisons
> and computations on the actual values can be performed. This lack of
> metadata standardisation is, in my mind, the weakness of FITS.
> 
> I also feel that creating a XSD which would be completely dynamic 
> and assign a specific tag to each parameter of each possible 
> astronomical table in the VO is currently hopeless; but Roy's
> suggestion of having a tool to transform VOTables into "hardXML"
> would indeed be interesting for applications making usage of 
> generic XML tools -- I imagine that the overhead of such conversions
> would not be important for "small" tables (and could be driven by
> XSLT scripts)
> 
> About XSD, I was surprised by Andreas' assertion:
> >BTW: As many people already pointed out there is no valid VOTable Schema as
> of
> >now, maybe we should first concentrate on trying to produce this. I think
> >everybody would agree that we should indeed have a Schema file which can be
> used
> >by standard parsers and XSD tools.
> 
> whereas in his following message 23minutes later, he gives an excerpt of
> the VOTable schema (which can be found e.g. at
>     http://vizier.u-strasbg.fr/xml/VOTable.xsd
> ) 
> Andreas, did I misunderstand anything ?
> 
> I expect also to be able to circulate the new VOTable1.1 definition
> document quite soon -- hoping it won't generate another avalanche...
> 
> --Francois
> ================================================================================
> Francois Ochsenbein       ------       Observatoire Astronomique de
> Strasbourg
>    11, rue de l'Universite F-67000 STRASBOURG       Phone: +33-(0)390 24 24
> 29
> Email: francois at astro.u-strasbg.fr   (France)         Fax: +33-(0)390 24 24
> 17
> ================================================================================
>