- From: Jonathan Marsh <jmarsh@microsoft.com>
- Date: Mon, 4 Nov 2002 15:19:02 -0800
- To: <www-ws-desc@w3.org>
Schema Normalized Universal Names (draft at [1]) is soon to be published as a W3C Note. In reviewing this document, I first notice that Schema NUNs are relative to a particular schema. They are not relative to a target namespace. The Use Cases given don't seem to make this distinction; that is, NUNs can be used by RDF statements to talk about a particular schema, but there is no explicit use case of talking more abstractly about a target namespace. Which do we want to identify for WSDL? Schema NUNs are compatible with XPointer. A new fragment scheme xs-nun() uses an XPath-like syntax to distinguish the various symbol spaces and navigate the hierarchy of schema components (just as XPath distinguishes node types and names and hierarchy). Here is an example of a schema NUN (on multiple lines for readability): http://www.example.com/schemas/po.xsd# xmlns(po=http://www.example.com/PO1) xs-nun(/complexType(po:Items)/sequence()/item) Note that the syntax identifies constructs relative to a QName in the "complexType" symbol space. A corresponding WSDL identification mechanism might look something like: http://www.example.com/wsdl/myticketagent.wsdl# xmlns(ta=http://www.airline.wsdl/ticketagent/) wsdl(/portType(ta:TicketAgent)/ operation(listFlights)/ input(listFlightsRequest)) Arthur's proposal looks more like this: urn:wsdl:http://airline.wsdl/ticketagent/# input(TicketAgent/listFlights/listFlightsRequest) These are just syntax differences. [1] http://lists.w3.org/Archives/Member/w3c-xml-schema-ig/2002Oct/att-0050/0 1-part
Received on Monday, 4 November 2002 18:19:42 UTC