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

RE: RE: Issue 146 proposed resolution

From: Henrik Frystyk Nielsen <henrikn@microsoft.com>
Date: Thu, 15 Nov 2001 09:17:21 -0800
Message-ID: <79107D208BA38C45A4E45F62673A434D05801943@red-msg-07.redmond.corp.microsoft.com>
To: "Doug Davis" <dug@us.ibm.com>, <xml-dist-app@w3.org>

I buy that. However, while the base SOAP 1.2 processing model doesn't
address this directly I think it is important to enable this kind of
scenario as part of SOAP based systems. The term *sender* wasn't in this
case intended to be seen as a raw SOAP node but rather a part of an
application with a notion of the intended or ultimate destination of the

Henrik Frystyk Nielsen

>>In this case, the party who the sender thinks is the ultimate 
>I don't disagree with the rest of your note, but this one line 
>does concern me.  To some people using the "actor" attribute 
>gives the sender some sort of routing or targetting mechanism. 
>We know that routing is definitely not what it is there for, 
>and targetting is only true so far as you can say "the node 
>that will process this block will act in 'this' role".  So, 
>when we talk about a sender "thinking" that a particular node 
>is the ultimate destination it sounds a little too close to 
>targetting a node by something like an IP address rather than 
>its role. A sender only has the notion of two things (w.r.t. 
>targetting or routing), and that is "next" (since he's the one 
>who opens up the socket) and "everything else" - which 
>technically could wind up being "next" depending on the roles 
>"next" chooses to take on.  I get a little concerned when I 
>see text that seems to imply the sender has any more control 
>or knowledge beyond this.  So, I don't believe the sender can 
>ever really "think" a certain party(node) is the ultimate 
>destination - he must always look at "next" as sort of a black 
>box - "next" could be the end of the road or could be just the 
>beginning - its a total unknown.
Received on Thursday, 15 November 2001 12:19:06 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 22:01:16 UTC