W3C home > Mailing lists > Public > public-ws-async-tf@w3.org > March 2005

Re: Ueber-MEPs and points North

From: David Hull <dmh@tibco.com>
Date: Mon, 28 Mar 2005 17:11:53 -0500
Cc: public-ws-async-tf@w3.org
Message-id: <42488129.2000700@tibco.com>

This doesn't look so bad, actually.  A question:

Without referring to HTTP, how would you characterize the case when the 
reply address is the default, the fault address is non-default and a 
fault occurs?

Having answered that, what happens from the point of view of HTTP, and 
how does the abstract SOAP MEP map to that?

Ideally, this would generalize to multiple reply addresses and multiple 
fault addresses, with an arbitrary subset of them actually carrying 
messages as part of the MEP.

Yalcinalp, Umit wrote:

>My recollection is that the "in" segment is not optional, although there was some mutterings about it which I did not quite catch, probably due to the fact that we have also out* MEPs that we need to deal with. 
>My understanding is the following chart with respect to the relationships between the WSDL MEPs and the uber SOAP mep (in-optional-out):
>WSDL in ==> SOAP in-optional-out MEP. The optional output message is never generated, but the http response would be generated to be 202. 
>WSDL robust-in => SOAP in-optional-out MEP. The output message will contain the SOAP fault only. 
>WSDL in-optional-out => SOAP in-optional-out MEP. (Works only when non-anonymous ReplyTo is not used). 
>AFAIK, we have touched upon, but did not agree whether WSDL in-out may yield two separate SOAP MEPs when ws-addressing is engaged, but not in the last concall. 
>Any comments/corrections are welcome. 
>-----Original Message-----
>From: public-ws-async-tf-request@w3.org [mailto:public-ws-async-tf-request@w3.org] 
>Sent: Monday, Mar 28, 2005 1:50 PM
>To: public-ws-async-tf@w3.org
>Subject: Ueber-MEPs and points North
>For the benefit of those who, through nobody's fault but their own, 
>didn't make the last conference call (and for anyone else just reading 
>the list and not on the concalls), could someone outline how the new 
>"über-MEP" would work?  From context, I gather that it would be composed 
>of an "in" segment and an "out" segment, both optional, with "in-only", 
>"out-only" and "in-out" falling out as special cases.  Is this basically 
Received on Monday, 28 March 2005 22:12:25 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 19:48:42 UTC