So, I did some research, and answered my own question. Section 3.3.4.2 of "Definitive XML Schema" by Walmsley, says that the schema below is illegal as the name "bar" will be interpreted as a reference to xsd:bar, not the targetNamespace. However, it seems many XML Schema processors may be tolerant of this error. On Fri, Mar 16, 2012 at 4:48 PM, Mike Beckerle <mbeckerle.dfdl@gmail.com> wrote:
Suman,
In and XMLSchema/DFDLSchema do I have to qualify the names of types?
We have a bunch of test schemas written roughly like the example below. They all have the default unprefixed namespace as XML Schema's namespace. They also all have a target namespace.
But some or all of the type references to named types use unqualified names. In my mind, that means they would be assumed to be in the XML Schema namespace, not the targetNamespace.
On the other hand, the XML Schema validator doesn't complain. But that just means the schema is valid, not necessarily meaningful.
Example here:
<schema xmlns="http://www.ogf.org/dfdl/dfdl-1.0/XMLSchemaSubset" targetNamespace="http://example.com"> <element name="foo" type="bar"/><!-- IS THIS LEGAL, no prefix on name of the type. --> <complexType name="bar"> <sequence/> </complexType> </schema>
Mike Beckerle | OGF DFDL WG Co-Chair Tel: 781-330-0412
-- Mike Beckerle | OGF DFDL WG Co-Chair Tel: 781-330-0412