W3C home > Mailing lists > Public > xml-dist-app@w3.org > November 2001

RE: Usage Scenarios

From: John Ibbotson <john_ibbotson@uk.ibm.com>
Date: Thu, 15 Nov 2001 09:47:02 +0000
To: "Nilo Mitra (EMX)" <Nilo.Mitra@am1.ericsson.se>
Cc: xml-dist-app@w3.org
Message-ID: <OFA8DB15F8.0EA4A8F2-ON80256B05.0031A1AD@portsmouth.uk.ibm.com>

Nilo,
Comments below. I have added a number of example messages to the document
which should provide more illustrations. As promised on the call yesterday,
I'll make sure the draft is "complete" before the F2F.
John

XML Technology and Messaging,
IBM UK Ltd, Hursley Park,
Winchester, SO21 2JN

Tel: (work) +44 (0)1962 815188        (home) +44 (0)1722 781271
Fax: +44 (0)1962 816898
Notes Id: John Ibbotson/UK/IBM
email: john_ibbotson@uk.ibm.com



                                                                                                                  
                    "Nilo Mitra (EMX)"                                                                            
                    <Nilo.Mitra@am1.er       To:     John Ibbotson/UK/IBM@IBMGB                                   
                    icsson.se>               cc:     xml-dist-app@w3.org                                          
                                             Subject:     RE: Usage Scenarios                                     
                    11/14/2001 09:23                                                                              
                    PM                                                                                            
                    Please respond to                                                                             
                    "Nilo Mitra (EMX)"                                                                            
                                                                                                                  
                                                                                                                  






John:
Here are some comments on the usage scenario descriptions ([1]):
1. Should we not make a global change from "XMLP" to "SOAP", or "SOAP 1.2"?
<JBI> Agreed - changed XMLP to SOAP in text and all diagrams

2. Section 2.1.2, last sentence: This is the *only* place where
"XMLP_UNITDATA.send" is used. I would delete it, as it is does not really
add much to the discussion and you don't want to get caught up explaining
the additional terminology.
<JBI> Agreed - I've deleted the last sentence

2bis. Section 2.4.2., first para: This talks of "XMLP Handlers". I'm not
sure this is a defined concept, and the text following about the handlers
being required to do the parameter and result serialization seem too
prescriptive.
<JBI> I've changed this section in line with the RPC examples where the
Body is used for the request arguments and result

3. Text immediately following Fig. 5: "For the RPC programming model, an
RPC Request handler on the XMLP Sender serializes the calling parameters
and places them in an XMLP Header in the calling request."  This seems too
prescriptive.
<JBI> Replaced with example RPC request/response messages

4. Same place, the very next sentence: " The XMLP Receiver has a matching
XMLP Handler which extracts the serialized parameters and invokes the
procedure (which is local to the XMLP Receiver).". Can one be sure of the
statement in parentheses? Is that statement necessary?
<JBI> Replaced and simplified with examples

5. Section 2.5.2. Why is a transport supporting the request/response
assumed? This could equally well have been provided by two unidirectional
messages. Is it necessary to have a status block generated for the
request/response scenario? For example, a HTTP POST can be acknowledged
with a "202" or "204" status code (and no response body) could serve
equally well for this scenario.
<JBI> Scenario S5 specifically requires an acknowledgement request of some
kind and I've interpreted this as being independednt of any underlying
transport ack. Reliable delivery and date/time of delivery is the example
I've used. I agree that two unidirectional messages could be a solution,
but that is the case for all of the scenarios so in this case I simply
assumed an HTTP type of transport - I don't then have to add a correlation
header.

I'll provide others as I go through it some more later. Better to get this
partial review out to you sooner rather than be complete - and late.

Best regards
Nilo

[1] http://www.w3.org/2000/xp/Group/1/10/18-us/XMLProtocolUsageScenarios

Nilo Mitra
Ericsson Internet Applications Inc.
phone: +1 516-677-1073
mobile: +1 516-476-7427
email: nilo.mitra@ericsson.com

   -----Original Message-----
   From:   John Ibbotson [SMTP:john_ibbotson@uk.ibm.com]
   Sent:   Friday, October 19, 2001 10:18 AM
   To:     xml-dist-app@w3.org
   Subject:        Usage Scenarios

   The latest draft of the usage scenario descriptions is available from
   http://www.w3.org/2000/xp/Group/1/10/18-us/XMLProtocolUsageScenarios
   Please can you review it and provide feedback. The remaining scenarios
   will
   be added asap.
   John

   XML Technology and Messaging,
   IBM UK Ltd, Hursley Park,
   Winchester, SO21 2JN

   Tel: (work) +44 (0)1962 815188        (home) +44 (0)1722 781271
   Fax: +44 (0)1962 816898
   Notes Id: John Ibbotson/UK/IBM
   email: john_ibbotson@uk.ibm.com
Received on Thursday, 15 November 2001 04:56:57 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:59:04 GMT