RE: ANNOUNCE - SchemaWeb

Victor Lindesay said:

> SchemaWeb is at: 
> Your feedback is welcome. 

Thanks for this. It is very nice indeed, and a very useful service.

Apologies if this seems like nit-picking, but I have one query in that
the display of metadata about a schema displays an entry "namespace",
and I'm not quite clear what "namespace" means in this context. Is it
the XML Namespace Name typically used in RDF/XML? Or is it used as a
synonym for "vocabulary"?

Anyway the data available suggests to me that there is a one-to-one
correspondence between a schema and a "namespace". 

Patrick Stickler argued long and hard (and to me, at least,
convincingly) on this list, see e.g.

and elsewhere that namespaces-as-XML-Namespaces are simply part of the
"punctuation" of the RDF/XML serialisation of RDF, and that there is no
necessary one-to-one correspondence between:

- what the RDF specs call a vocabulary (a set of URIrefs) (I think there
was an RDF Core decision to use the term "vocabulary" rather than
"namespace" for this purpose precisely in order to avoid any confusion
with XML Namespaces), 
- a schema (an RDF/XML document providing info about the resources
denoted by those URIrefs), and
- an XML Namespace. 

It _may_ be that for all my classes and properties with URIrefs
beginning, I create RDFS descriptions in one
RDF/XML document which I make available at that URL, but there is no
_requirement_ that that is the case.

In fact one of the examples indexed by the registry highlights this. The
metadata for the schema for the RDFS vocabulary



But if you click on "classes and properties", you see that the data
provided in the schema currently located at

does (quite reasonably) include descriptions of resources such as etc - terms "from
another namespace", if you like, which of course are also described by
the schema located at

and described by 

So I think I'm wondering whether describing a relationship between a
schema and an XML Namespace is perhaps inappropriate; and while the
relationship between schema and vocabulary (if that is what is intended
by "Namespace" in this case rather than XML Namespace) may be worth
describing, it is potentially many to many, as this example suggests?



Received on Wednesday, 19 November 2003 09:28:57 UTC