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

Re: Issue: Problem with ProcedureNotPresent fault subcode

From: <noah_mendelsohn@us.ibm.com>
Date: Fri, 3 May 2002 10:39:46 -0400
To: Jacek Kopecky <jacek@systinet.com>
Cc: moreau@crf.canon.fr, skw@hplb.hpl.hp.com, xml-dist-app@w3.org
Message-ID: <OF6AB42251.A1791FC9-ON85256BAE.005201D6@lotus.com>
The case where you do recognize the QName is easy, for the reason you 
state.  I think the questionnable case is:  the QName is a representation 
of the procedure name.  So a QName comes in.  Your code does the 
following:

1) Is it one of the QNames I process as a known body document type -- NO
2) Does it correspond to one of the methods I process as RPC - NO

OK, which fault do I return.

------------------------------------------------------------------
Noah Mendelsohn                              Voice: 1-617-693-4036
IBM Corporation                                Fax: 1-617-693-8676
One Rogers Street
Cambridge, MA 02142
------------------------------------------------------------------







Jacek Kopecky <jacek@systinet.com>
05/03/02 10:35 AM

 
        To:     noah_mendelsohn@us.ibm.com
        cc:     moreau@crf.canon.fr, <skw@hplb.hpl.hp.com>, <xml-dist-app@w3.org>
        Subject:        Re: Issue: Problem with ProcedureNotPresent fault subcode

 Noah,
 thanks for this info. I think that in this case we must decide 
how we understand it in SOAP 1.2:
 a) first you understand the first QName in the Body, then you 
know it's RPC and the ProcedureNotPresent fault has no meaning 
then (and also it seems that RPC cannot forbid multiple 
invocations in one Body then),
 b) or RPC is application of SOAP governing the whole Body, in 
which case ProcedureNotPresent has sense.
 We are currently saying we don't mandate any kind of semantics 
in Body (except for the Fault), which would indicate that we'd 
better stick with (b) as then RPC gets to say what Body is (as 
opposed to some higher level of contract which uses RPC for 
particular QNames).
 Best regards,

                   Jacek Kopecky

                   Senior Architect, Systinet (formerly Idoox)
                   http://www.systinet.com/



On Fri, 3 May 2002 noah_mendelsohn@us.ibm.com wrote:

 > History:  FWIW, I was involved in consideration of this issue for the 
 > design of 1.1.  There, we explicitly decided that a receiver was 
expected 
 > to understand the QName in the body, and from that would understand any 

 > conventions associated with the use of that body.  I agree with Henrik 
 > that it's circular (and unnecessary) to model a misunderstood Body 
QName 
 > differently for RPC than for any other use of the Body.  Not 
recognizing 
 > that it's a purchase order (presumed message style) is no different 
than 
 > not recognizing getStockQuote (presumed RPC style), IMO.  Only when you 

 > recognize it do you understand whether there are good mappings to the 
RPC 
 > method/args style (and as Henrik points out, the degree to which you 
 > actually do a nice binding to a programming language is completely 
private 
 > to your endpoint.)
 > 
 > ------------------------------------------------------------------
 > Noah Mendelsohn                              Voice: 1-617-693-4036
 > IBM Corporation                                Fax: 1-617-693-8676
 > One Rogers Street
 > Cambridge, MA 02142
 > ------------------------------------------------------------------
 > 
 > 
 > 
Received on Friday, 3 May 2002 10:59:14 GMT

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