Is schemaLocation Required When Importing Inline Schemas?

The spec currently says that when you refer to inline schemas within the 
same document you need to provide a schemaLocation attribute with a 
suitable fragid in the xs:import statement:

Inside an embedded XML schema, the xs:import and xs:include element 
information items MAY be used to refer to other XML schemas embedded in 
the same WSDL description, provided that an appropriate value is specified 
for their schemaLocation attribute information items, such as a fragment 
identifier (see [XML Schema: Structures] 4.3.1). The semantics of such 
element information items are governed solely by the XML Schema 
specification [XML Schema: Structures].

I thought we said the fragid was unnecessary since the WSDL processor 
knows where to find the schemas.

On the other hand, there was a very long thread [1] about how to refer to 
an inline schema from another WSDL document. The thread ended 
inconclusively with an action item [2] that doesn't seem to exist. The 
conclusion seems to have been that you didn't need the schemaLocation, but 
the behavioir was dependent on the processor.

Can we please tighten this up? I just created some Primer text that may 
not be correct.

I propose that we require a WSDL processor to know how to locate any 
inline schemas it encouters while processing a document so that the 
schemaLocations are not required.

"If schemaLocation is not present then a WSDL processor MUST attempt to 
locate the schema among the inline schemas it has encountered while 
processing the WSDL document."

[1] http://lists.w3.org/Archives/Public/www-ws-desc/2003Nov/0109.html 
[2] http://lists.w3.org/Archives/Public/www-ws-desc/2003Nov/0156.html

Arthur Ryman,
Rational Desktop Tools Development

phone: +1-905-413-3077, TL 969-3077
assistant: +1-905-413-2411, TL 969-2411
fax: +1-905-413-4920, TL 969-4920
mobile: +1-416-939-5063, text: 4169395063@fido.ca
intranet: http://labweb.torolab.ibm.com/DRY6/