"Rick Jelliffe" <ricko@[...].au>  wrote:
| From: "Arjun Ray" <aray@[...].net> 

|>  Right.  Namespaces are neither necessary nor sufficient for the general
|>  problem of vocabulary combination.
| 
| I think this is where the new generation schemas languages fit in. 

They could, but I wouldn't be so quick to th conclusion that schema
languages are needed here.

| The namespaces set general semantics, 

I don't think this is true.  I would attach the notion of semantics to
"vocabulary" (in the sense that it's been bandied about in discussions of
this sort - see the Namespaces Rec too!)  rather than to "namespace" (in
the sense of the Rec).  The fact that a bunch of universal names share a
URI "prefix" is no more than a coincidence.  They need not be parts of any
coherent schema at all.  They're just a smorgasbord.

Conversely, the notion of "vocabulary" does carry an implication of an
underlying schema - that there are definite ways to organize the names
when applied to a document.  If it's meaningful to say that a document
could have only a single vocabulary, then the possibility of multiple
vocabularies raises the issue of knowing (or determining) to which parts
of the document each of these would apply.  Which in turn would suggest
the legitimacy of the notion of a vocabulary-specific "view", i.e. the
part of the document which is *coherent* with respect to a particular
vocabulary.  (Contrast this with random collections of allegedly universal
names - there is no necessary notion that they contribute to a coherent
whole in terms of their *own* "namespace".)

| the schemas tell you how they must be combined. 

But this is not necessarily an issue of schema combination.  There's a
difference between whether schemas are mergeable and whether they are
miscible - for the particular document instance only.  That is, the use of
multiple vocabularies could have been a one-off - there was no intent to
synthesize a new schema (as a repeatedly instantiable document type).

IOW, it should be possible for a document, as an instance by itself, to
claim conformance to different schemas in its various parts, without being
obliged to assert that there is some unitary schema for the entire
document unifying these various schemas.

But that's still jumping the gun.  The minimum necessary is a means for a
document to assert what parts of it are "assigned" (by authorial intent)
to a vocabulary specific view, such that if anyone cared they could run a
validator on that view to see if it comformed to the corresponding schema.
(We *know* that's a separate issue: this discussion is about parsing, not
validation.)

| What neither of them necessarily set is the semantics of combination (i.e. 
| what do these information items mean together, what do they mean apart)  
| The best that a schema can do is constraint a document language to only 
| accept documents that has certain semantics-of-combination. 

Yes, but before one checks against a schema, one must know what part of
the document to submit to the check.  I don't see why every schema must
apply to all of the document, or why a document should have a unitary
schema when all the author intends is conformance of various parts only.


-----------------------------------------------------------------
The xml-dev list is sponsored by XML.org <http://www.xml.org> , an
initiative of OASIS <http://www.oasis-open.org> 

The list archives are at http://lists.xml.org/archives/xml-dev/

To subscribe or unsubscribe from this list use the subscription
manager: <http://lists.xml.org/ob/adm.pl>