OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: URIs and information typing

Eric van der Vlist wrote:
> Jonathan Borden wrote:
> >
> > right, so: "xsd:unsignedInt" is just shorthand for
> > http://www.w3.org/2000/10/XMLSchema#unsignedInt -- which is references a
> > location in the XML Schema schema <simpleType id="unsignedInt">... ?
> If I am not mistaken, the association is done though the id attribute
> that has been added to the simpleType element in order to make sure that
> the fragment id "http://www.w3.org/2000/10/XMLSchema#unsignedInt" points
> to the corresponding element of the schema.
> This is working fine with schemas having a stable location that is
> pointed by this namespace and which element names are duplicated by
> their ids,  such as http://www.w3.org/2000/10/XMLSchema, however, the
> fragment identifier "http://example.com/ns/example#mytype" would not
> point to the definition of "myns:mytype" if the schema location was
> anything else than "http://example.com/ns/example" or if the id of the
> simpleType element wouldn't match its id.

this is a big problem. it is unacceptable to have different W3C recs having
completely different and, on reading CR XML Schema-1 20001024, arbitrary
rules for mapping XML QNames to [namespace name, localname] pairs or to URI.
The rules for constructing a text/xml or application/xml fragment id are
indicated by RFC 3023 as XPointer, and (IMHO) it is essential that a QName
have a direct and unambiguous interpretation as a URI. XML Schema datatypes
are (again IMHO) otherwise broken.