W3C home > Mailing lists > Public > xml-dist-app@w3.org > September 2002

RE: Proposal for issue 306: Is use of Appendix A optional?

From: Jacek Kopecky <jacek@systinet.com>
Date: 06 Sep 2002 17:42:48 +0200
To: noah_mendelsohn@us.ibm.com
Cc: Martin Gudgin <mgudgin@microsoft.com>, Henrik Frystyk Nielsen <henrikn@microsoft.com>, XMLP Dist App <xml-dist-app@w3.org>
Message-Id: <1031326968.3124.18.camel@krava>

 Noah,
 I think you may have a point, therefore now I can see it equally both
ways - moving the mention in 2.1.1 or keeping it in RPC. Using SHOULD
could be OK, as I said there's no interop problems involved (because the
XML form is the interface) and a MUST would be unenforcable here.
 Best regards 

                   Jacek Kopecky

                   Senior Architect, Systinet Corporation
                   http://www.systinet.com/


On Fri, 2002-09-06 at 01:06, noah_mendelsohn@us.ibm.com wrote:
> 
> I think I agree with what I take to be Jacek's position.  2.1.1 is fine is 
> written prior to this discussion.  Edge names >are< Qnames in the model, 
> therefore cannot raise the problem of being difficult to serialize.
> 
> Now, in the case of RPC the situation is different.  There we explicitly 
> state that our purpose is to provide a means of representing method names 
> and named arguments that often originate in some (unspecified) external 
> system.  It is those external method and argument names that potentially 
> violate the rules of a QName, therefore I think it's appropriate that the 
> reference to appendix B go in the RPC section.  I am fairly strongly 
> opposed to putting it in 2.1.1, as that seems incoherent (for the reason 
> above).
> 
> Regarding MAY vs. MUST, I think the answer might be SHOULD.
> 
> <existing 4.2.1>
> The invocation is represented by a single struct or array containing an 
> outbound edge for each [in] or [in/out] parameter. The struct or array is 
> named identically to the procedure or method name (see B. Mapping 
> Application Defined Names to XML Names).
> 
> Each outbound edge either has a label corresponding to the name of the 
> parameter (see B. Mapping Application Defined Names to XML Names) or a 
> position corresponding to the position of the parameter.
> </existing 4.2.1>
> <proposed>
> The invocation is represented by a single struct or array containing an 
> outbound edge for each [in] or [in/out] parameter. The struct or array is 
> named identically to the procedure or method name (the conventions of B. 
> Mapping Application Defined Names to XML Names SHOULD be used to represent 
> method names that are not legal XML names.).
> 
> Each outbound edge either has a label corresponding to the name of the 
> parameter (the conventions of B. Mapping Application Defined Names to XML 
> Names SHOULD be used to represent parameter names that are not legal XML 
> names) or a position corresponding to the position of the parameter.
> </proposed>
> 
> ------------------------------------------------------------------
> Noah Mendelsohn                              Voice: 1-617-693-4036
> IBM Corporation                                Fax: 1-617-693-8676
> One Rogers Street
> Cambridge, MA 02142
> ------------------------------------------------------------------
Received on Friday, 6 September 2002 11:42:51 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:59:11 GMT