VOTable alternative?

[Guy Rixon]

On Mon, 19 Jan 2004, Tony Linde wrote:

> > > Ultimately, VOTable works because it can describe external, binary
> > > formats that are never addressable by XPath.  If we replace
> > VOTable,
> > > then we need to retain this abaility.
> >
> > Unfortunately, I'd have to agree with this one...
>
> I thought FITS did this (and here I am really out of my depth). Wouldn't it
> be easier to stick with FITS for binary data (or enhance it if it doesn't
> describe metadata adequately) and XML Schema / XML for documents in xml
> format?

FITS metadata are inadequate for the VO.  The standard doesn't constrain the
use of metadata significantly, so there are many "valid" FITS-files which are
semantically useless.  Most of them, in fact.  This is exactly why VOTable was
devised.

It is very useful to be able to describe XML tables and binary tables in the
same metadata format.  It is very useful that the metadata format is XML and
thus can be processes using the "XML standard tools".

It is much less useful that the data themselves be processable by "the
standard XML tools". In general, the powerful tools will die screaming when
then see a 10GB table because they use DOM.  Even if they scale to arvitrary
lengths of XML data, they still don't do binary formats, so aren't a general
solution.

I think having a metadata format in XML for all table formats is excellent,
even if VOtable 1.0 isn't the right one.  I would rather protect that value by
throwing out TABLEDATA altogether than by using one kind of metadata for XML
and another for binary.