VOSI: capabilities

[Ray Plante]

Hi GWSers,

I've been looking at how we can tighten both the VOSI and TAP 
specifications to clarify exactly what a TAP service should return for its 
capability metadata (S2.1 in VOSI, 
http://www.ivoa.net/Documents/VOSI/20090825/WD-VOSI-1.0-20090825.html). It 
seems to me that we probably need to be more specific about the 
representation of this metadata.

The form of the capability metadata is currently described in this 
paragraph:

   The service metadata shall be represented as an XML document which
   contains a sequence of one or more elements of type
   {http://www.ivoa.net/xml/VOResource/v1.0}Capability or sub-types
   thereof.

I see a few issues with this:

   1.  I believe the intent here is that the XML document has an arbitrary
       root element containing direct child elements (as a sequence) of
       type vr:Capability; however, it is actually ambiguous.  It seems
       that the sequence could have any number of ancestors (i.e. elements
       between it and the root element) and still be compliant with this
       statement.

   2.  Even if we have eliminated the ambiguity of (1) above, I think we
       need to provide an explicit schema to use for this document (as
       VOResource is not sufficient).  This schema would define a root
       element that takes a sequence of elements of type Capability.
       Without this, either implementors would have to provide their own
       custom schemas that do the same (and post them on-line) or we would
       be faced with responses that are not validate-able using standard
       XML Schema techniques.

To address these, we need to first be more explicit about the XML layout. 
If provide the schema, then doing so is easier: we say, must be compliant 
with XXX schema with the root element xxx:xxx.

Assuming we do provide a schema, the question becomes how.  The document 
already provides a simple schema for the availability metadata (namespace 
http://www.ivoa.net/xml/Availability/v0.4).  We could add a new global 
element for capabilities to this schema, but we should probably change the 
namespace name (to, say, http://www.ivoa.net/xml/VOSI/v1.0). If we have 
some entrenchment with http://www.ivoa.net/xml/Availability/v0.4 (which I 
doubt we do), then we would need a second schema.  Any preferences?  May I 
propose a new schema?

cheers,
Ray