FW: reftype (IDREF constraints)

[Arnold, Curt]

Robin Cover wrote:
>> So: are any of the new specs addressing the
>> matter of constraints upon referenced elements?
>> In current syntax, this would amount to saying
>> that IDREF(s) in a given case could only be
>> pointers to specified (classes of) GIs, as
>> approximately in the HyTime 'reftype'.

The schema datatypes document mentioned that whether there should be
constraining facets on a uri was an issue.  It is not a big jump to go a
constraint on the schema to also allow constraints on the target.

>From the Working Draft, section 3.2.14
 Issue (uri-scheme-facet): should we have a facet to allow a limitation to a
specific scheme?  It might be useful to able to say that something was not
only a URI, but that it was a "mailto" and not a "http://...". 

>> Where would/should this concern for scope
>> (semantics) be addressed in the W3C drafts?
>> Are there any dramatic proposals in the works
>> for bedrock re-design, or improvement of the
>> ID - IDREF mechanism?

I'm thinking schema datatypes would be a good place for part of it, if it
should be done.  You could allow the definition of facet on a datatype
derived from uri or idref to constrain the possible targets and then use the
datatype on the linking attribute.

Something like:

<schema>
	<datatype name="uriToSomeElement">
		<basetype name="uri"/>
		<!-- some representation of a 
			constraining facet on the datatype, just making this
up  -->
		<constraint>
			<elementTypeRef name="SomeElement"/>
		</constraint>
	</datatype>

	<elementType name="SomeElement">
		<attrDecl name="id">
			<datatypeRef name="id"/>
		</attrDecl>
	</elementType>

	<elementType name="OtherElement">
		<attrDecl name="someElement">
			<datatypeRef name="uriToSomeElement"/>
		</attrDecl>
	</elementType>
</schema>

In the case of URIs, you could only express an intention as to what is on
the other side of the URI.  On an IDREF, the processor could determine if it
matched the facets.  The facets would have to allow a match to one or more
elementType's in the current schema or declared namespaces, one or more
elementTypes defined in a specific DTD or ad-hoc elements, mime types (if it
is a uri to a JPEG for instance), maybe others.  Again, I'm not sure it
should be done, but that is where I think it might fit best.

I'm really interested in seeing what the W3C Schema group has in store for
their next draft (as I'm sure most everybody on this list is.)

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)